Home Store Products Research Design Strategy Support News
 News January, 2011

Designer’s Blog: January, 2011

31 January, 2011

In moving Stavka forward, one of the issues I will need to address is on-map play aids. One reason I have been delaying this is that I haven’t been certain what play aids the game would need. My current thinking is that it will need the following:

Of these, only the first two are really informational displays; the others are really nothing but blank spaces on the board that help the players keep their stuff organized. Finding map space that is available for use for play aids can be a problem — it was a big problem in The Guns of Gettysburg because with the moving objectives it was never clear where the battle was going to take place — but I think that Stavka has enough suitable areas that it shouldn’t be a problem. Stavka play aids in principal can go in Scandinavia/The Baltic, Italy/Yugoslavia, The Black Sea, and the Urals without getting in the way of the game.

Of the various play aids, the time track is the most flexible in terms of layout: it can be one long row, or it can be coiled up in various numbers of rows and columns as needed. It even has a fair amount of flexibility in terms of the sizes of the boxes for the individual turns. The map legend has some layout flexibility, but it needs to be big enough for full-size terrain samples and labels, and for aesthetic reasons it needs to have an aspect ration broadly around 4:3. The reduced armor areas also have some flexibility, although they also should have a roughly 4:3 aspect ratio and need to be large enough to hold perhaps a half-dozen armor pieces comfortably.

The areas with the least amount of flexibility are the card areas (draw deck, discard pile, and armor pool) since these need to be the size of the cards used in the game, plus a small decorative margin. Of course, I could go with playing cards with custom sizes or dimensions, but I would much prefer to stay with standard-sized playing cards.

Another question regarding the play aids has been language. I have made the map bilingual (Russian/German) but have always assumed that text in the play aids would be English. I have not, however, actually tried it out and am a little concerned that it might look odd. If it does look odd, I have grave doubts about the wisdom of making Russian/German play aids, and even graver doubts about Russian/German/English play aids, and so I fear that I will have to choose between accepting its oddity and abandoning my bilingual map, which I don’t want to do. Still, I suppose I will cross that bridge when I come to it.

30 January, 2011

In the Stavka turn structure, after the production phase, the front line is reset to reflect penetrations made in the previous turn. Either side is also allowed to make voluntary withdrawals by abandoning cities. Cities that were front line cities in the previous turn can be abandoned at no penalty, while those that are not front-line cities can be abandoned only at a penalty of one card per city. That’s the idea anyway.

There is a potential problem in that system concerning which side can abandon cities. In theory, the offensive player (such as the Germans in the early part of the game) could withdraw to force the creation of Soviet salients, which they could then cut off. This is undesirable. My current thinking is that a player cannot abandon a city if the nearest penetration to that city is theirs. If a city is equidistant to penetrations of both players, neither player can abandon it. This seems sensible, but I’m not sure as yet whether it is sufficient. In particular, if neither side made a penetration in the previous turn, then neither side can abandon any cities. That might be ok, but I can’t say for sure at this point.

One other activity occurs during this phase is that each player is allowed to voluntarily remove any of his strong-points from the map. The timing of this, after the production phase, is intended to complicate the defense and work to the advantage of the attacker. The reason this works is that the cost of a strong-point depends on the number on the map, but because in the turn sequence old strong-points can’t be removed until after new ones are built, the old strong-points increase the cost of the new. In early versions of the game, this rule wasn’t in effect, and it produced a weird form of virtual “teleportation” of strong-points where a player could first delete his old strong-points and then cheaply build replacements in the desired locations, more quickly and flexibly than even armor could be moved. It was weird and bad, but the new rule neatly eliminates it at no real cost in complexity.

29 January, 2011

Another cyberboard playtest of the 2nd edition of Bonaparte at Marengo wrapped up, and there was another French victory. After taking it in the chin in earlier playtests, the French have definitely come back. Still no decision on whether to try the V13 revisions discussed yesterday though. I am, however, leaning more towards giving them a try; it isn’t like I can’t back them out if I don’t like them.

Stavka also had another playtest, this one much more successful than the previous one. I did retain a couple positive features out of the previous train-wreak playtest, so even that wasn’t a total loss. I didn’t pursue this test far, only into the winter of 1941-42, but it was never my intent to go any farther. I think that Stavka’s problems are, one by one, slowly getting resolved, but I’m really feeling the need for some sort of mathematical baseline model to help me arrive at the right magic numbers.

To give you some better idea how the game works at present, each turn starts with a production phase, in which the Soviets and Germans draw new cards and make payments for things. In the current design, the list of things that can be (or must be) paid for in the production phase is as follows:

As you can see, there is not that much to production. There is one wrinkle that I haven’t mentioned, and that is that leadership cards count double in strong-point builds, armor rebuilds, and armor builds. (I have been going back and forth, incidentally on whether the leadership multiplier should be limited to one card per purchase; if so, I should definitely cut the armor build price from ten points to eight.)

There are still multiple unresolved issues in production, in addition to those noted above: There are no clear rules for ordering of actions between the Soviets and the Germans, it isn’t clear whether the definition of a deep penetration for resupply should depend on where it starts or where it ends, and it isn’t clear what the limit numbers should be on cards placed in the armor build pool. Still, in spite of these issues, things are undoubtedly getting clearer. Sorting these things out just takes time.

28 January, 2011

I thought I’d talk a little bit about the design of the locales near the Austrian entry area in the 2nd edition of Bonaparte at Marengo. Below you can see illustrations of different versions of them (the version numbers are the numbers assigned during playtest):

Version 11 Version 12
Version 13 (?)

Version 11 is the original design I made when I first drew the map. Version 12 is the most recently tested version, and version 13 is one that has not been tested, but which I am considering as an option for future testing. (The large locale id numbers, incidentally, will not appear on the production map.) The difference between 11 and 12 is that locale 3 in version 11 has been split into two locales in version 12, numbered 3 and 4. The difference between versions 12 and 13 is that locale 4’s eastern boundary has been changed; in version 12, locales 4 and 9 are not adjacent, but in version 13 they are.

So, what is motivating these various changes?

First of all, all locale sizes generally represent the degree to which the terrain is obstructed by terrain, slowing movement. Where the terrain is open, the locales are generally large, and where it is difficult or closed, the locales are small. Boundaries correspond to natural terrain features as much as possible, given the map scale. Because of the number of significant terrain features in this part of the map, most of the locale boundaries are easy choices, forced by terrain features.

If you look at version 11, locale 3 is quite large, but it is almost entirely surrounded by difficult terrain. Generally, I try to avoid such a large locale where this is the case. What is more, in version 11 locale 3 is very long on the north-south axis, quite a bit longer than I generally allow. However, locale 3 has no obvious internal boundary along which a dividing line would go, which is why in the original version of the map, it was left as a single large locale.

Version 12 tries a split in locale 3, breaking it up into two locales. From a terrain modeling perspective, I think version 12 is probably better than version 11, and what is more, I think it promotes somewhat more interesting game play in the opening, because it makes the Austrians make decisions earlier about which forces are going to press north and which south, and it slows the Austrians up just a touch. Still, it isn’t a very big change in the overall design of the game.

Version 13, on the other hand, would be a more substantial change. In fact, unless there were other changes in the game, it would probably stress the French defense to the breaking point. The key difference between versions 12 and 13 is that it would increase the minimum number of approaches that a French forward defense would have to cover from 2 to 3. With the two pieces just off-map to the east and one initial activation, that would give the French 3 pieces to defend 3 approaches, which just isn’t viable, especially since it would leave them with no reserve to cover the road to Castelceriolo. Unless the French got very lucky in their initial piece draw, they would almost immediately be unable to defend not only the forward positions, but the Fontanone stream as well.

Now, from a historical point of view, there is nothing wrong with making it harder for the French to defend the forward positions: they historically did so only very briefly. However, a design that makes a French defense of the Fontanone almost impossible is historically quite dubious — in the actual battle it was the main French defensive line and they held all morning into the afternoon. Further, from a strictly terrain analysis point of view, version 12 is a better model of the terrain than 13 is.

So, given all this, why am I considering version 13 at all?

Well, in part because I think that the forward defense is probably easier in the game than it should be, given that the French were historically surprised by the Austrian attack, and what’s more, it isn’t a particularly interesting game play option. I think that the game would be better, both as history and as a game, if I made it harder and less attractive. The goal isn’t to make it impossible, as I think that the French could have fought longer than they historically did on that line if they had wanted to, just more difficult. I would like it to be more of a real decision point in the game, rather than an easy default line of play.

And so, version 13 comes basically from that motivation. Although it is objectively an inferior terrain analysis, divorced from historical context, as part of a general simulation of the entire battle, in context it could be part of a reasonable simulation. However, it can’t be made in isolation or it will wreck the Fontanone defense as an option, and so I am considering it in combination with another change that would allow the French to activate two pieces on the first turn instead of one. As a bonus (in theory, anyway) this would also make the initial game play more interesting for both sides. Those would all be good things, but what I’m not convinced of yet is that this will really do that, nor that it is the best way to do it. And so, it remains only a potential change.

27 January, 2011

Playtesting for both the 2nd edition of Bonaparte at Marengo and Stavka has been trundling along.

BaM2 has had a number of face-to-face games with French wins, so concerns about a pro-Austrian balance problem may have been calmed. One tester even thinks that it is the French that have the edge. I’m less concerned about balance, but I’m still not entirely happy with the opening game. I’ve made a small map adjustment (as well as correcting an unintentional road inefficiency introduced during an earlier locale change) in response, and am considering another. Other changes I’m thinking about are small morale level adjustments, and a revision to the objective locales. Nothing decided yet though. In general, there are no rules problems and the game is playing well, but I’d like it to play even better if I can think of a good way to do it.

The last Stavka solo playtest (the only way Stavka gets played, until it is further along) was a dismal mess. I had a number of new ideas for changes to the game, threw them all in at once, and the result was a complete train wreck. I did learn a couple of useful things from it though, and expect to run another test soon, with most of my “innovations” from the last test backed out. I think that I need to provide base production numbers per turn (rather than it always being the total of active production centers), with deductions for enemy-held production centers, and after a delay for conversion, additions for captured enemy production centers. It isn’t more complicated than a total, and I think it gives me the flexibility I need to handle the different phases of the Soviet war effort (initial mobilization vs. ongoing production). We’ll see how it goes.

26 January, 2011

The last several blog entries have been used to show you the various components of The Guns of Gettysburg being sent off for quotes by printers. This is the last of these entries, which show the box top and bottom.

Top Bottom
Gettysburg Box Top Gettysburg Box Bottom

(Click on images to open them in their own window.)

There is not much new here, except that this shows you what the files look like untrimmed. The four corner sections of the art files are not visible on the finished box, since they are where the paper folds around the box corners. There are also large margins on the edges of the art. Some of the margin will be wrapped around the box edges, while the excess is trimmed off.

I don’t expect to make many more posts about GoG from this point on, although you can expect some updates as the manufacturing process moves forward. The next major milestone will be when I get production samples, and at that point I will finish the web pages for the game (which are currently just a stub linking to the diary). After that, I will get a firm date and will start taking pre-orders. After that will come, you know, actually shipping actual games to customers.

25 January, 2011

Here are images of the files for the rules and map components of The Guns of Gettysburg. (The rules have had a number of small revisions, but the map hasn’t changed since it last appeared in this diary and is included here only for completeness sake.)

Rules Map
Gettysburg Rules Gettysburg Map

(Click on images to open them in their own window.)

The main area of recent work was the rules index, on the play aid page in the back of the rule book, which had numerous small errors that had crept into it since the last revision. (In theory, Adobe InDesign, the program in which the rules were written, supports automatic indexing, but I wasn’t happy with the automatic indexing results or the process.)

The play aid page on the back of the rule book is one of those things that I really should have always done. (Napoleon’s Triumph would particularly have benefitted from it.) You can see it below:

Gettysburg Rules Play Aid

One of the issues I had in designing it was understanding exactly what it was for. There is a lot of information that potentially could be on it, but what should be on it? As it worked out, it is a patchwork of different elements, each of which muscled its own way onto the page.

The first of these were the state flags, which are used on the game pieces. These found there way on it in part because they are attractive designs and add a splash of color to an otherwise gray page. (I greatly believe in trying to avoid any page of solid text in a rule book; rules by their nature tend to be intimidating, and images are a great help in softening that impression.) Also, frankly, the game needs a key to the flags somewhere, because who knows all the state flags anyway? Especially since these are antiquarian state flags, so even if you know the modern flags you wouldn’t necessarily know these.

Next were the orders of battle for the opposing armies, as represented in game pieces and tokens. I thought this was important information that the players should have without having to count through a large number of blocks and tokens to get it.

What information to extract and summarize from the rules was kind of a thorny question. At one time, I played with a “key rules” idea, but in practice I decided I didn’t know what a “key” rule was: there are lots of rules that, if omitted, tend to blow a hole in the game, far too many for the space available. After messing with that idea for a time, I decided on a different approach: that the rules summary of the play aid would be about procedures and modifiers: lists and numbers.

Finally, of course, there was the index. I always intended to include an index, but wasn’t sure how to best construct it. Ultimately, my decision was made by space limits. The index is about as much information as I could cram into the single column allocated for it, and is basically built around the rules’ technical terms — those terms introduced in bold-face and then formally defined in the rule book.

24 January, 2011

Another The Guns of Gettysburg component that is all spec and no art are the wooden pieces:

Gettysburg Wood Pieces

(Click on image to open it in its own window.)

There is not a lot to say about these. The dimensions and colors are the same as those used in my previous games. The biggest change is the switch from silkscreening to stickers, which is a tradeoff of one set of advantages for another. The main strengths of silkscreening are their durability and that the pieces are ready-to-use right out of the box. The main strengths of stickers are the higher resolution and color, which make it possible to make a much greater variety of piece designs than is possible with silkscreening. For Bonaparte at Marengo and Napoleon’s Triumph, the limited number and simplicity of the piece designs made silkscreening the better choice, but for GoG, with its more varied and more complex designs (largely owing to the command names on the pieces) stickers were the choice that made the most sense.

23 January, 2011

The game tray is a The Guns of Gettysburg component that is all spec and no art. You can see the 2-page spec below:

Page 1 Page 2
Gettysburg Tray Spec, Page 2 Gettysburg Tray Spec, Page 2

(Click on images to open them in their own window.)

I have some experience with metal game parts in that the command pieces in Napoleon’s Triumph were cast zinc. (Zinc was chosen because it flows well into small molds when cast.) It certainly would have been possible to have had cast metal trays, but the weight of a solid metal part of that size would have been prohibitive. (The games do need to be shipped, and shipping cost goes up with weight.) Another alternative would have been Scrabble-style wood trays, but I thought I’d go with a more industrial product as befits the period, and so I decided to go with stamped sheet metal. I haven’t had a price quote yet, but this is supposed to be the cheapest form of metal product, and so I am reasonably optimistic that the price will be acceptable.

For a material, I decided to go with steel. I considered aluminum, but aluminum is quite light, and the game trays need to be heavy enough that they won’t accidentally move when brushed by a player’s hand. The only way to have aluminum trays as heavy as I wanted would have been to use much thicker sheets, and the sharpness of the bend you can get with sheet metal is inversely proportional to the thickness of the sheet. (Try to bend sheet metal too sharply and it fractures.) Steel, however, rusts, and so I specified nickel-plating to protect it — a customer suggested that as a suitable low-cost rust-prevention process.

Each tray is designed to hold 8 tokens comfortably; a token is .8 inches wide, so 8 tokens close-packed are 6.4" wide. The trays are 7.5" to allow an inch of additional room. Originally, the tokens were .75" wide, but were increased to .8" in order to allow some additional margin for die-cutting error. When that was done, the tray width, which was originally 7" was increased to 7.5". The heights of the trays are high enough to support the tokens, but still leave a bit of the token poking over the top of the tray. This makes the tokens easier to pick up, and also allows players to easily see how many tokens their opponent has in his tray. (The number of tokens each player has is not supposed to be secret, only their identities.)

22 January, 2011

A couple more images of the files sent out in the quote request package for The Guns of Gettysburg. This time, the sticker sheet art and sticker die specification.

Art Die
Gettysburg Tokens (Front) Gettysburg Tokens (Die)

(Click on images to open them in their own window.)

One of the late changes I made to the physical design was to the objective markers. My purpose all along was to have markers that laid flat on the map and looked as if they were part of it. My live playtest games were all with printed acetate markers, which accomplished this goal very well, but there were, unfortunately, issues. The first was that they were so thin that they were very difficult to pick up. Second, they were so light that it was easy to sweep them to a different position on the board by accident. Finally, they got lost easily, as they were very difficult to see. All-in-all, I have to describe the experience as something of a failure, which was too bad, because I had a lot of interesting ideas as to how such components could be used in future games. But if they didn’t work, they didn’t work.

The only reason they stayed in the design as long as they did was because I was unsure about what to replace them with. I still wanted something that would like flat on the map, so blocks were out. Thin slices of wood were another idea, but wood as a material really doesn’t like to be too thin. I thought of adding them to the counter sheet and having them as cardboard, but a star shape would leave a lot of edge, and the exposed cardboard edges of wargame counters always have an unfinished look to them. (I was prepared to deal with the edge for the tokens, because there was no practical alternative material for them.) I also thought about clear plastic. Really, clear plastic was the obvious material for the job, but I had issues with the use of plastic in an ACW game. It just didn’t feel very period to me. I also thought of more exotic materials like metal, but frankly the expense of metal for the objectives just didn’t seem justified. Burying the metal under stickers made it seem so unnecessary — visually, there would be no sign that they were metal, and there was no structual quality of the pieces that required metal as a material. After going back and forth over the same stupid issue and the same alternatives literally for months, I gave in and decided to go with clear plastic winks, with star stickers on them. Not ideal, but then nothing was ideal — that was the problem. They’ll work fine, and I think they’ll look fine, but I have never been able to think of them with any fondness; they are very much something I felt forced to do rather than something I wanted to do.

On the plus side, I do think the sticker sheet looks quite pretty.

21 January, 2011

I mentioned in the previous blog entry that I had prepared the files for getting printer quotes. I thought I would show some of them. Below are images of the three files used to specify the die-cut tokens used in the game. The first two images are of the artwork files for the front and the back of the token sheet, while the third image is of the specification for the die used to cut them.

Front Back Die
Gettysburg Tokens (Front) Gettysburg Tokens (Back) Gettysburg Tokens (Die)

(Click on images to open them in their own window.)

The artwork files are slightly larger than the actual sheet. The extra color around the edges is called a bleed. It gives the printer a margin for error on his cut lines, without which there is the risk of a fine white line on one or more edges where the paper didn’t get any ink.

20 January, 2011

Well, I spent much of yesterday putting together a quote package for The Guns of Gettysburg. I don’t know how much folks know about how the publishing business works (some more than others, I’m sure), but publishers don’t typically own any of the equipment to print and manufacture games. That work is contracted out to printers or other manufacturers, depending on what is to be made. The first step in that process is to get quotes from potential printers, which in turn requires that they have enough information to make a quote. This information basically consists of all the art files for the game, and the manufacturing specifications for the various components.

Now, GoG was pretty much ready for me to get a quote, but there was still some final work to be done. My art files generally have layers that are useful to me while working on the game, but which aren’t useful to the printer, and in fact are potential sources of confusion. (For example, I GoG has a layer with locale id numbers, which I use in playtesting, but which I don’t want printed on the production map.) So, I make copies of the art files and delete all these non-production layers from them. Another important step is to do a final pass on the rules. The purpose of this pass is to check cross-references of various kinds, since these are the items that most easily get out of sync as the rules change during development. Such references include the table of contents, the index, play-aids, and page headers. Most of yesterday was spent doing final checks. (Of these, the index was the most time-consuming.) Finally, a master file is produced, which lists and describes all the game components and identifies which files are for which components. At that point, everything needed to get a quote request is completed.

Once complete, I upload the results to the server and send emails to printers pointing them to them and requesting a quote. And that is what I will do today. If, from this, it sounds like I am moving forward with publishing GoG, well, that is correct. I don’t know at this point when it will be ready. One of the things I’ll get from the printers, along with quotes, are schedules. However, as nothing ever goes perfectly, there are always some issues that come up that add time to the schedule, and I’m sure that will be the case here. Generally, I don’t announce a date until the games are actually being printed, which is the earliest point in the process that I have any confidence as to when the game will be ready. So, don’t ding me for any publication date. I will give a date when I have one myself.

19 January, 2011

One of the most important qualities I look for in a game design is tightness (as distinct from looseness). Tightness is a subjective quality in a game; it is when the player feels that his choices are difficult, but that it is important that he make the right choices in order to win. The opposite is looseness; a player’s subjective sense that his choices are easy and/or unimportant. Tightness contributes to a player’s sense of involvement and enjoyment of a game, while looseness induces apathy and boredom.

One major element that drives tightness vs. looseness is balance. In general, if a player feels that a game is balanced against him, the game will feel tighter to him, but only up to a point: as the balance becomes more and more unfavorable, eventually a tipping point will be reached where the player concludes that he can’t win no matter what he does. At this point, instead of feeling that his decisions are critical, he feels they are irrelevant; he can’t win no matter what he does and so the game feels completely loose. While imbalance against a player induces tightness (up to the tipping point), imbalance in a player’s favor always induces a sense of looseness. Therefore, for both players to experience a game as being tight, it is necessary that the game be reasonably balanced.

Balance is important to creating a tight game, but it isn’t sufficient in and of itself. Players not only need to be unsure whether they’re going to win, they need to be unsure how they’re going to win. To really have a tight game, the game has to make the players have to sweat over what they’re going to do. One of the great things about game design is that there are an awful lot of ways to get this particular quality. Chess and Poker, for example, both have it, but they achieve it in completely different ways. For Chess, the problem is trying to see the future, to look ahead into a complex tree of possible futures that gets overwhelmingly large as it gets more remote. The tension comes from the concentration required and the intellectual difficulty of the task. For Poker, the test is more of judgment of character and of nerve rather than abstract intellect: Do you know your opponent better than he knows you? Can you control your nerves while breaking his? What Chess and Poker both have in common, though, is that the players feel that it is their decisions that decide the game, but that the right decision isn’t easy to see.

The process of making player decisions more difficult, more challenging, is what I call tightening the game. Ideally, players should feel that the decisions are hard right from the beginning and all the way to the end, although it is quite common (and not unacceptable) for there to be a couple loose turns at the beginning, if the players are starting from pre-determined situations. Initial looseness is not fatal, providing that the game gets tighter as it goes along. It is also common for a game to lose tightness as it goes for the player who is gaining the upper hand during the game. To some extent this is inevitable, particularly if the players are not closely matched, but the main thing here is to avoid a prolonged but pointless end-game where the game keeps going even after the outcome will generally be obvious. (That was a fault in the 1st edition of Bonaparte at Marengo, which allowed games to continue even after one side was demoralized, an event that almost always effectively ends the game and should have formally ended the game as well.)

I think that one of the most successful aspects of Napoleon’s Triumph is that it is a very tight game. (One of the least successful is the learning curve of the attack resolution procedure, but that’s a whole different area of game design.) The tension kicks in right away with the initial set-up and continues as each player’s plan unfolds and they adapt to what their opponent is doing. That morale losses only apply to the loser of an attack help to keep the tension up during play, since the sense is there that a big win can reverse even a bad situation — and indeed it often can. The game’s major kicker is the French reinforcement decision, which can make the whole game turn around, but it is also a very tough decision that abounds in negative as well as positive consequences.

Looking ahead, I think that The Guns of Gettysburg is a nice, tight game, but probably less so than NT. (The balance between offense and defense had shifted so between the two periods being simulated, that big attacks in the Civil War were much more likely to break the attacker than the defender, which tends to make players more cautious.) GoG does have some advantages, though, even compared with NT. The unpredictability of the reinforcements is challenging, and the complexity of the terrain on the battlefield compared to that of NT (Gettysburg was a much less open battlefield than Austerlitz) rewards close analysis and punishes carelessness. The variable turn length and general order system are also nice tightening design features not found in NT.

With regard to the 2nd edition of BaM, making it a tighter game is partly a matter of correcting the balance, and partly a matter of having the game tend more to build towards a climax, rather than tending to gradually loosen, as the 1st edition of BaM1 was prone to doing. It is tricky work, for a variety of reasons, but I think BaM2 is already tighter than BaM1, and I’m still looking for opportunities to make it tighter yet.

I admit I’m not spending a lot of time just yet worrying about Stavka’s tightness. The basic game concept is sound, I think, and will make for a good game, but it is far from being a finished system. My major concern at present is that my plan to carry the game through the end of the war runs the risk of having a game that keeps going after the outcome has been decided (which as mentioned above is generally a bad thing), but I have some ideas on how to address it, and I have always thought it a pity for the Soviets in eastern front games to have to take a lot of punches early and never get a chance to really return the favor. So, we’ll see.

18 January, 2011

Although I’ve gotten a reputation for producing a certain kind of game, I actually don’t have much in the way of fixed beliefs about what makes a good game, mostly because I think there are a lot of different ways for a game to be good. One thing I do believe in though, is in trying to make the kind of games that I enjoy playing myself. I won’t start development on a game if I don’t think I would enjoy playing it, and I won't close development on a game and publish it until and unless I do enjoy it.

Apart from that most basic approach though, there are certain things that are characteristic of my taste in games. One of them is a preference for small numbers and consequently very basic arithmetic. If you look at some of the number ranges in my games, for example, you can see this:

Bonaparte at Marengo (1st ed.)1-30-21-3
Napoleon’s Triumph1-30-11-3
Guns of Gettysburg1-20-21-2 (hourly)*
Bonaparte at Marengo (2nd ed.)1-30-11-4
Stavka (as of now)1-30-21*
* Feature unlimited distance “strategic” movement (no counting)

The main reason I like smaller numbers is that people can calculate with them quickly, and because wargames tend to run long, I”m always looking for ways to speed play. Another reason is that because my numbers are small, they also are in a very narrow range, and even the smallest change (+1 or -1) generally makes an important difference. Since, to me, arithmetic is a negative element in the game experience, I want there to be a payoff for any arithmetic operation I make a player perform. If I make him add one to a total, I want that plus one to matter; it shouldn’t just be a waste of his time and effort. That which doesn’t matter shouldn’t be in the game at all.

17 January, 2011

The use of front-line pieces in Stavka is in large part an aesthetic decision, but it also reflects one of the most important design elements of Stavka: the abstraction of infantry. Strategic east front games in general are notable for the large number of infantry units they include, almost all of which are basically front-line filler; their job is to create the front line separating the opposing armies. In Stavka, all of this front-line-creating infantry is represented by the front line they create. From a cause and effect perspective, the game shows the effect by abstracting the cause. This is essentially an inversion of the way traditional games on this subject work, which abstract the effect but show the cause.

I owe this idea to this picture, submitted by Rusty Ballinger to the website BoardGameGeek, where he drew the front line onto a photo of a game picture he took:

Looking at it, I realized that the image of the front line was more visually revealing about the state of the game than the picture of the pieces themselves, and I thought that if it was more revealing, then that is what the game should have. My first thought was to use a chain in the game to mark the front line, but I quickly realized that it wouldn’t work once I began to think about how a chain would actually work in play: there was no way to easily alter just one section of the front line without pulling on the chain and causing the entire front line to move around.

With this realization, it became clear that what I needed instead was a set of parts that could be linked and unlinked easily, so that one part of the front could be left in place while another part was moved. The other part was that the links had to be able to connect at any angle, since the front had to be able to curve and undulate from one end of the map to the other. After playing with a lot of different alternatives, the best approach I came up with was the hook and loop design. Another aspect of the design is that the hook is wider at the opening and tapers towards the top. By making the hook wider at the opening, less accuracy is needed when dropping it down onto a loop, while the taper towards the top makes for a snugger fit once it finally drops down. Ideally, the top of the hook is the same width as the metal used to make the piece, but the test piece (as shown) has a wider gap because it is cheaper to make, and I don’t want to spend more money than I need to on test pieces.

Another aspect of the design is that the pieces are intended to be stamped from sheet metal rather than molded, which is a cheaper process. (And with a lot of metal pieces, I do need to watch the cost if the game is going to be reasonably priced.) I am concerned, with this design, however, about the loop on the back and how it is to be machined. It is a critical element to this design, and if it can’t be machined reasonably cheaply, then I’ve got a problem. But we’ll see.

I’ve gone ahead and ordered some playtest front-line pieces from eMachineShop. Here’s hoping that they come out well!

16 January, 2011

Sorry to have gone dark for the last few days on the blog. I’ve been traveling and wasn’t able to post any updates. Just before I left though I found an interesting site called eMachineShop. They have a specialized (and simple) CAD program you can download, and then use it to specify what you want machined, upload it to them, tell them how many you want, and then they’ll machine them and send them to you. (Assuming they are able to do so — they try to make sure that what you design in their app is something that they can machine, but their tests aren’t infallible.) Now that is pretty cool. Also pretty cool is that they have a 3D viewer so you can see the part you’re designing. I tried it out with the front-line pieces I’ve been working on for Stavka, and you can see what one of them is supposed to look like below:

The picture is huge, but these parts are pretty small: the piece as a whole is 1.25" long, with the end loop being .25" in diamter. Now, physically, the way these things work is that they have a hook on one end and a loop on the other, and in the game the front line is constructing by hooking these things together to make a chain. A turn of Stavka is basically a series of attacks by each side to penetrate the front line, and then at the end of the turn the front line is reset to reflect the results of those attacks. Resetting the front line means unhooking the parts of the front line that have been penetrated and replacing them with new sections that reflect the results of the penetration. You can see a series of conceptual illustrations below:

Front Line at Turn Start
Operations (Attacks)
Reset of Front Line at Turn End

Now, territorial control of the game is in terms of cities, and the front line runs between them and indicates which cities are Soviet-controlled and which cities are German-controlled. The gaps between the cities are fairly large (a couple of inches) so the game isn’t physically fussy about placement. An inch one way or another makes no difference as long as the line runs between the right cities, and the number of links in the chain doesn’t matter, so when playing, just go with what’s easy — or, with whatever looks cool, though easy often tends to produce pretty cool-looking fronts without any particular effort.

On another note, I think I’ve figured out how to manage supply within the framework of my cards/production system. Initial tests have been positive, but the real test is when I try to actually to calibrate the “magic numbers” of per-turn card income for each side.

12 January, 2011

Well, the 30 gauge aluminum sheet metal has arrived. It is in a roll, which I would really rather it wasn’t, but I haven’t found any non-rolled aluminum of the desired thickness. It certainly seems thin enough to work and cut, with proper tools, but as I have neither I shall need to do a little shopping. I am not confident that I can work it with enough skill to make it aesthetically superior to the polyclay pieces I already have, but I will be giving it a try.

I have not yet done another Stavka playtest. I continue to mull over the way the card system models production, and I think the core problem is that while it models production, it doesn’t really model supply as a separate constraint; that is, the tendency of an advance to outrun its supply. I had thought that the costs paid for penetration movement would do the trick, but as currently implemented, I am coming to the conclusion that they don’t. I hate (very much) to introduce a new sub-system into this game and so I’m trying to figure out how the current system can be modified to handle supply better. I have a few ideas I’m playing with, but nothing yet that seems worth trying out.

Playtesting of the 2nd edition of Bonaparte at Marengo continues to percolate along. It is pretty much all-Cyberboard at present. Now Cyberboard has its pluses and minuses. Its best feature (and it is a very, very good feature) is that it automatically generates records of each playtest game that can be subjected to later analysis. Very helpful feature that. But, its worst feature is that by its nature it tends to be very, very slow. Still, with GoG still ahead of BaM2 in the production pipe, time is not critical for BaM2, so the slowness of the process is not a big problem.

One issue that I continue to ponder is what I should do next after Stavka. I have a number of ideas, but have not settled on any of them. Before I begin work on a game, I like to at least have a mental model of how it is generally supposed to work. Lacking that, the concept sits in what the movie business refers to as development hell, although in the game business it would be better called conceptual hell. (In the movie business, development means the opposite of what it does in the games business. In the movie business, development means that the project isn’t moving forward, while in the game business it means that it is.) With Stavka development stalled for the time being, and BaM2 creeping slowly forward in playtest, requiring little time from me, it would be nice to have something else to work on, just so this time isn’t wasted. Wanting something, however, doesn’t make it happen. One of these concepts needs a breakthrough idea so that I know how to go forward with it and that hasn’t happened yet. Oh well. At some point I’ll figure out how to do one of them. I do hope that it is sooner rather than later though.

11 January, 2011

I thought I would share with you some shots from the first Cyberboard playtest of The Guns of Gettysburg, back in 2009:

5PM, 1 July 11AM, 2 July
6AM, 3 July 12PM, 3 July
1PM, 3 July 7PM, 3 July

In the first day of the battle, not much happened. The players were completely new to the game, and when you don’t understand things well, your reaction is not to attempt much. It wasn’t until late in the first day that the first attack occurred. The main event of the second day was a violent attack at 11:00AM by the Confederates to eject the Union from McPherson’s Ridge, an attack that was successful, but with heavy Confederate losses. The rest of the second day was characterized by bringing up reinforcements and extending the lines to the south. In the early morning of the third day, the Confederates attempted a major attack on their far right, which failed. The Confederates then shifted their attention to the Union center on Seminary Ridge, which they attack and drove off at 1PM. They then tried to swing south and make a series of attacks into the gap that had opened up between the Union left and Center. At first, these attacks were successful, but there was not enough fresh Confederate infantry to keep pressing forward, and artillery ammunition was extremely low. The Confederates tried to regroup and made one last small-scale attack at at 7PM. When that failed, the Confederate player resigned.

From a simulation point of view, the game at this point worked quite well. The game encouraged historically reasonable tactics and gave historically reasonable results. And (from a simulation point of view) no substantial rules problems were discovered. However, in terms of learnability, it was clear that the game had major issues. From the email exchanges before and during play it was painfully obvious that a lot of work was ahead before players could easily grasp what was going on. Now, the general tendency in games is that they add rules over time: a problem is found, a rule is added to fix it. Another problem is found, another rule is added and so on. To take a game late in its development and move it the other way, removing rules and simplifying, is exactly the reverse of the natural process and is just painful work. Every rule that you remove was added for a reason, and if you just remove it, all you do is restore that problem that caused the rule to be written in the first place. Some problems you have to just decide to live with: that fixing them just isn’t worth the complexity required by the solution. Other problems you decide that you can’t live with and you need to find another, simpler way of solving. However that “simpler” solution will often have problems of its own that aren’t revealed until testing, and fixing those tends to mean adding more rules, and before you know it, all the complexity you thought you removed has crept back in.

This whole process is what I will always (un)fondly remember as the GoG playtest death-march. It went on forever. In the first period, it was mostly about trying to simplify the game without breaking anything. In the second period, it was mostly about trying to get the balance right, which proved an extremely difficult problem due to the variable reinforcement schedule. It really wasn’t until we had a large backlog of Cyberboard test games, when I could take all the games and apply a computer model to them, that I was able to come up with rules that generally balanced the game. There are still some statistical outliers that remain unbalanced, but overall, I think GoG (finally) came out of it all as a well-balanced game, and that the quest to improve playability did not significantly compromise the historical accuracy that had been achieved earlier on. I sure would like future development projects to be easier though.

10 January, 2011

The nickel sheet metal (30 and 28 gauge) arrived today. I might be able to work it, but I think I will wait for the aluminum sheets to arrive. Aluminum is supposed to easier and I think easier is still better from where I am.

I (finally) finished reading The Wages of Destruction, an economic history of Nazi Germany. It was an altogether interesting read, and if it was short on the sort of specifics I was hoping for (the extent to which Germany was able to make use of the resources of the occupied Soviet Union) it did give me a general idea. I think that one of my core ideas for the game, to allow the Germans to gradually make use of Soviet production centers is historically reasonable. Certainly without such an idea, the entire rationale of the German invasion of the Soviet Union, and indeed the objectives they set for their military campaigns, is unintelligible.

I did another playtest of Stavka, with a higher initial German card count and a lower per-turn draw (at least in 1941) and I think it felt closer to the historical campaign, but the way that the production cards mesh with operations still doesn’t feel quite right. I’m not sure if tweaks to my current framework can give me the results I want or if I need to make some more fundamental changes. As of yet, I’m not even sure I can really articulate what is wrong about it, which certainly has a great deal to do with why I’m not yet sure how to fix it.

9 January, 2011

Well, bummer. My sheet metal didn’t come yesterday after all. Darn post office. Still, I suppose that one way or another I will manage to cope with the disappointment of getting my sheet metal two days late (by the normal human calendar) or one day late (by the USPS calendar).

I edited the various line of sight blog entries together and put them in the design section of the sight as an essay. (Which you can read, if you’re so inclined, here.) Doing so was always my intent, assuming that the material came out well, and I think it did. The essay I did on chance in wargaming still gets referenced quite a bit. I doubt the line of sight essay will get similar treatment. As I discovered when I published Bonaparte at Marengo, chance in wargames is something that flares up in occasional controversy, and on such occasions, someone or other will frequently drop a link to my chance essay. I don’t think the same thing happens with line-of-sight though. Still, I am pleased with the line of sight essay (and as soon as I say this, I know that someone will find an editing error in it), both in form and content. Even if it doesn’t get a wide readership, this is the sort of thing I enjoy thinking about and it is the sort of problem I enjoy working on.

I continue to ponder the balance issues in the 2nd edition of BaM. The key problem is that although the French won the historical battle, the situation actually favored the Austrians. They had a bigger army (slightly), the advantage of surprise, the advantage of concentration, they had more guns, and (important for a battlefield like Marengo) more cavalry. Really, the most likely outcome was an Austrian victory. And in fact, the Austrians did win the “first” battle of Marengo in the morning through the early afternoon, only to lose the “second battle“ in the early evening. Even the French thought they were fortunate to escape with a victory.

Now, the 1st edition of BaM dealt with this by requiring the Austrians not just to win the “first” battle by forcing the Fontanone, but to aggressively move forward and take distant objectives. This was, I think, reasonable in historical and game terms. The Austrian strategic situation was pretty dire: the French had crossed the Alps into their rear and had cut them off from Austria. The battle of Marengo was an Austrian attempt to break out and escape back to Austrian-controlled eastern Italy. The fact that BaM1 generally favored the French was the product of the rules making a French delaying action easier than it should have been, rather than any inherent French advantage in the battle.

However, in correcting this historical inaccuracy I’ve exposed the inherent imbalance in the historical situation. Even after all the French reinforcements arrive, their army is not likely to be bigger than the Austrian. One corrective I’ve applied is to give the French a higher morale. Now, this too is historically reasonable. It was the Austrians, after all, that cracked in the “second” battle and the French who held on and prevailed. There are limits, however, to much I can adjust this and still feel comfortable with it historically, and also how much I can adjust it without risking odd lines of play resulting from the high morale imbalance.

In any case, I don’t think the balance issue is actually very great. This tends to rule out most changes to the core systems of movement and combat, which would tend to produce large swings in balance and could have other undesirable side-effects in the bargain. One thing I’ve been examining are the French activation rules. Now, historically, simulating the French initial surprise is a really “soft” area of the game. It isn’t clear how quickly the French could have reacted, for any given possible Austrian moves, nor how such reaction is best represented in the game. In BaM1, the surprise rules were fundamentally defined in game terms: they were designed to make it impossible to trap the Austrians at the narrow point of their bridgehead (which the French failed to do historically) but to make it possible for the French to make a prolonged defense of the Fontanone (which they did historically). These historical constraints apply to any French activation rules, but beyond that, it is partly a matter of what is simple from a rules perspective and what is interesting from a game perspective. (Yes, where history gives no definite answers, I feel quite free to let the needs of the game determine the answers.)

And so, I continue to consider options, while awaiting more playtest results. When in doubt, collect more data.

8 January, 2011

The most recent playtest of the 2nd edition of Bonaparte at Marengo has concluded, with French demoralization at 2:00PM. The picture below shows the game state immediately prior to the last two assaults that would have demoralized the French. (The testers didn’t actually execute the assaults once they both agreed that the outcome was Austrian victory.)

The French attempted to hold the Fontanone line as long as possible, giving up in the south around noon, but still tried to hold the northern end of the line even after pulling back in the south, and it was the decision to hold in the north that proved to be their undoing. The Austrians used their advantage in morale and elite infantry to hammer the northern end of the line, tipping the French over into demoralization while the Austrians with one morale point to spare. (Incidentally, the big locale ID's were just for the playtest map — they won’t appear on the production map.)

The game was close (a final morale difference of 1-0 can’t really get any closer), but I suspect that the game is more unforgiving of the French than the Austrians. This is actually the opposite of the general balance of the 1st edition, where most experienced players found the Austrians the less forgiving side to play. So, I think the balance is close, but I suspect it isn’t there yet. In terms of the game system, the game works well, and the system changes for the 2nd edition (which mostly involve wide approaches) add to the game, which was the general idea.

I am optimistic that BaM will get there. It is possible (even if I am skeptical of it) that it is there already, and that the next round of playtests will produce French victories even if I don’t change a thing. If balance were not an issue, there are actually a few small changes I would make that would favor the Austrians, but since I suspect that the game does not need ANY pro-Austrian tweaks, I will hold off on those — possibly forever.

With regard to Stavka I’m expecting some more sheet metal to arrive today. I ordered some aluminum sheets and some more (thinner) nickel sheets. They came from different venders though and it is the nickel sheets that I’m expecting today. One aspect of the Stavka piece situation is that I may find myself where I was with the 1st edition of BaM, where I hand-made test pieces. They worked fine in testing, but they didn’t look good enough to be used in box art and so the BaM box had no actual photos of the game in play. For Napoleon’s Triumph and for The Guns of Gettysburg I had good enough block pieces to use for box art, and for NT the pewter pieces I had made were usable (with Photoshop touch-up) in box art, but my hand-made counters for GoG were not good enough, which is one reason why they don’t appear on the GoG box. Getting pieces I can use for box art in Stavka is definitely going to be a challenge.

(Incidentally, if you have a copy of NT and compare the box picture of the command pieces to the actual pieces that came with your game, you can see the difference pretty easily. The pieces used in the box art were pewter, while the production pieces were zinc. The production command pieces are quite a bit darker and the metal was thinner: my supplier of the playtest pewter pieces could not make metal parts as thin as called for by the production specification. Much of my Photoshop touch-up of the pewter pieces was to thin out the metal so as to make them look more like the production pieces would look. I did not, however alter the color. I knew that zinc was a grayish metal, but didn’t know exactly how dark the pieces would be.)

7 January, 2011

This is my third (and final) entry in my personal history of playing east front strategic games. Panzergruppe Guderian:

PGG differed from the other entries, in that it is technically an operational rather than strategic scale game, but as I said about including Battle for Germany, it’s my list and I’ll put whatever games on it that I want to.

Prior to the publication of PGG, SPI had been stamping out game after game, most of them rather ho-hum, based on the Kursk game system. Now, the key characteristics of the Kursk game system were the move-attack-move sequence of play, odds-based combat resolution, plus soft zones of control. Odd as it may seem, given the large number of games based on it, I don’t think the Kursk game system ever actually worked very well. At first glance, PGG could in fact have easily fit into that series, as it had all the system’s key elements. It had the same basic sequence of play, an odds-based CRT, and soft zones of control. In spite of that, it played very differently, was wildly popular and went through multiple editions. So, what made it work?

The most obvious difference was that it had untried units whose strength was double-blind. (Even the Soviet player didn’t know what the strength of the pieces was.) Certainly this had a huge effect on one of the main aspects of the way Kursk-system (and even the oldest Avalon Hill) wargames worked: odds-tuning. Tactically, those games were built heavily around seeking magic numbers in attack odds. However, with untried units, neither side had the information needed to make those calculations. Both sides had to just throw units into attack and defense and hope for the best.

To understand what this might mean, let’s take a look at part of the CRT used in War in the East. (Actually WitE had 4 CRTs; this was the table most favorable to the attacker, used by the Germans in 1941.)

1 AeAeAeAExAExBrBr
2 AeAeAExAExBrDrEx
3 AeAExAExBrDrExEx
4 AeAExBrBrDrEx1/2Ex
5 AExBrBrDrEx1/2Ex1/2Ex
6 BrBrDrDrEx1/2ExDe

Now, here are the statistical loss ratios for both sides at the various odds, assuming the attacker has optimal pieces in the attack (no more points than are needed, with exact “change” for exchanges), with numbers expressed as a percentage of the defender’s strength, both in a normal situation and where the defender is surrounded:

Defender Surrounded42:17%66:33%50:50%33:67%50:87%50:100%50:100%

Unless the attacker can surround the defender, this is a pretty tough table for the attacker (and it isn’t even that kind if the defender is surrounded). And bear in mind, this is for an optimized attack. If the attack wasn’t optimized and the attacker didn’t have the right strengths to trade in an exchange, this table becomes just plain brutal. A 6-strength piece attacking a 1-strength piece, for example, might attack at 6:1, but because any exchange would cost the attacker 6 strength points, the loss ratio expressed as a percentage of the defender’s strength would be 402:87%, a statistical disaster for the attacker.

Now, it might at first appear that double-blind untried units were just as much a problem for the defender as attacker, but with a WitE-style CRT, that would not be the case. In fact, the lack of information favors the defender much more than the attacker, for several reasons. First, the table is exchange-heavy, but the attacker doesn’t know what is the optimal strength piece he will need for exchange fodder, greatly increasingly the likelihood that his losses will be heavier than the statistical norm because he will have to lose a higher-strength piece than if he knew what strength he would need for an exchange. Second, there is a subtle second-order bias in favor of the defender. Consider. Against a 6-point attack, if the defender’s strength is 1, then he may loss statistically 83% of his force, but it is 83% of 1 point, so it is only 0.83 strength points; no big deal. On the other hand, if the defender’s strength is 6, then the attacker will lose 66% of the defender’s strength, but it is 66% of 6 points, so he will lose 4 points. In short, the uncertainty might appear to affect both sides equally, but with a table like this it actually is much worse for the attacker than it is for the defender.

Given all this, you might think that untried units would require a very different CRT than the one used in WitE if attacking was to be possible at all, and you’d be right. Here is the same section of the CRT for PGG:

1 D1/A1D1D2D2D2D2De/A1
2 EngD1/A1D1D2D2D2D2
3 A1D1/A1D1/A1D1D2/A1D2D2
4 A1EngD1/A1D1/A1D1D2/A1D2
5 A2A1EngD1/A1D1/A1D1D2/A1
6 AeA2A1EngD1/A1D1/A1D1

Now, I can’t present the same sort of statistical losses for PGG’s table as for WitE, because the PGG table allows the defender and attacker a choice of losses and retreats. A D1, for example, allows the defender the choice of retreating one hex or losing one step. In his discussion of the game’s development in SPI’s Wargame Design book, Berg is pleased with this feature, and with good reason. It has any number of advantages over the CRT in WitE. First, as is necessary for a game with untried units, it tilts the balance more in favor of the attacker, which greatly helps the statistical problem noted in the previous paragraph, where a strong defender hurts the attacker far more than a weak defender hurts the defender, and makes the game much more forgiving of attacks against an unexpectedly strong opponent. Second, the CRT expresses losses in steps rather than as strength points, eliminating the attacker’s problem of providing the correct change for exchange fodder. Third, it largely avoids the problem of the WitE CRT, where depending on the situation a “worse” Dr result can be more desirable than a “better” Ex result, negating the motivation for seeking better odds. Fourth, and this is most important, by allowing the players to choose between taking losses and giving strength, it introduces actual strategy to applying CRT results, giving players in the game the same sort of choice their historical counter-parts had.

After untried units and a new CRT, the next interesting difference was the much more inviting overrun attack rules. Now, WitE had overrun rules as well: All you needed was 13:1 odds (!), using only one stack (and when overrunning, the German stacking limit was reduced from 4 to 3) and you could eliminate the defender during movement. Clearly, in WitE overruns were an extreme case in the game system, designed to prevent the Soviets from using a screen of 1-strength units to slow down large and powerful German forces, but otherwise not intended to occur. In PGG, on the other hand, overruns had no minimum odds requirement, but were resolved as normal attacks, but at 1/2 attack strength. Given the attacker-friendly and forgiving CRT, overruns were not an outlying case in PGG but an important part of the game.

Finally, the Soviet supply rules and the way their forces came into play prevented them from using the standard defensive strategy of the Kursk games, which was to build a solid wall of units from one end of the board to the other by the shortest possible path. Soviet supply tended to force them to build arcs of units to deny some approach paths, while at the same time leaving other approach paths largely open. The defensive object was to channel and trap the German advance. This created a highly fluid situation where both armies were constantly in motion, rather than grinding against each other along a static front line until one side or the other had its front penetrated. PGG was the first really mobile game at this scale I had ever played, and I think that was true not just for me, but for many other people. While much of its success was specific to the nature of this particular battle and not easily transferred to other games (SPI, like movie studios, was always on the lookout for a formula that would guarantee success), PGG still broke the stranglehold of the unloved and unlovable Kursk system on this period and scale, and that was a very good thing indeed.

If one was to look forward from PGG, I thought at the time that the logical direction for going forward was to take the success of the game’s overrun rules and go farther; eliminate the combat phase altogether as a redundant and unnecessary feature. I actually tried that, not in an armored warfare setting but in a tactical Napoleonic one; that was the underlying system for Thunder at Luetzen, which I published with SPI over 25 years ago. Was TaL a success? Only partly. I think the idea of eliminating the combat phase was sound enough, but not all of the decisions in the game were good ones. For example, the relatively bloodless CRT I also borrowed from PGG was a poor choice for Napoleonic infantry combat, even if it was perfectly reasonable for Napoleonic cavalry. Also, the game I think would have been improved by allowing cavalry to move through infantry zones of control, forcing denser and more realistic infantry deployments. Ah, well. Whatever was done with TaL was done. (And done long ago.)

6 January, 2011

A common observation about wargames is that they have a long playing time. There is, I think, a threshold that games cross when they can no longer be played in a single evening or a single afternoon. In general, we expect our entertainments to fit within such a time frame: we go out for an evening to a movie, we spend an afternoon watching a football game, and so on. This is just how people (at least in modern western societies) block their time.

Now, it is perfectly possible to design a game for an afternoon-evening double-block of time, with a break for dinner, but clearly the play opportunities for a game that requires such double-blocks are radically fewer than for a game requiring only a single block of time. Generally, double-blocks are available only on weekends, and there are usually other claims on people’s weekend time, making such double-blocks scarce. As an alternative, it is possible to play a longer game across multiple, non-consecutive time blocks: starting on one evening, and resuming on some later evening. Non-consecutive time-blocks, however, tend to run into space constraints: the game has to be left set-up somewhere, and just as there are other claims on people’s time, so there are other claims on people’s space. It is simply inconvenient to leave a game set up, and that inconvenience increases dramatically if more than one game is to be left set-up between sessions, so a game broken up that way tends to monopolize people’s gaming space, making it hard to play anything else until the long game is finished.

What all this means is that in terms of playability (a term often more associated with complexity than playing time), a game that can be played in a single afternoon or evening is highly desirable.

Now, I don’t think that wargame designers as a population are unaware of this, so why are so many wargames so long? Well, first of all, I think that once a problem reaches a certain level of prevalence in a population, it becomes accepted as normal. Since so many wargames are so long, wargame designers come to accept long playing times as normal. Designers come to think that that’s just the way wargames are, so there is no sense fretting about it. And then designers move on to other problems. (And there are always lots of other problems.)

But saying that wargames are long because designers accept that wargames are long is almost a tautology. What we want to get it is to try to identify the qualities that MAKE them long in the first place.

Now, there are two things that can take time in playing a game: thinking and acting. If you watch a game of Chess, a great deal of time is spent thinking, but not much time acting: players might spend 5 minutes deciding on a move, and then make the move in 2 seconds. Wargames, although they have the reputation of being very cerebral games, are not like cerebral in the same way that Chess is cerebral. If you watch a typical wargame, the players are much more physically active than Chess players. They may pause to just think on occasion, but mostly they are doing things: moving pieces, rolling dice, declaring attacks, doing table or rule look-ups, putting down or taking up markers, updating record tracks, and so on. Wargames aren’t slow because they require so much thought, they are slow because they require so much physical activity.

The physical activity a game requires can be partly understood as a product of the number of per-piece operations the game multiplied by the number of pieces. All things being otherwise equal, increasing or decreasing the number of pieces or operations will proportionately increase or decrease playing time. More subtly, playing time can be decreased by the mechanical complexity or simplicity of the operations themselves. One of the oldest, and therefore least recognized, mechanical efficiencies in wargaming is the use of fixed movement costs to determine how far a piece can move, rather than a die roll. (This was so unusual among games that the oldest Avalon Hill games used to actually reassure players in their rule books that dice were not used for movement in order to avoid player confusion.) The elimination of dice for movement was a huge time-saver, and probably cut the playing time of early wargames by 75% compared to hypothetical versions of those same games that used die rolls to determine how far each piece could move. It meant the difference between a three-hour game and an twelve-hour game, between a successful game and a commercial failure. In effect, it is this savings in mechanical efficiency that made wargaming as a hobby possible. I can’t see how the hobby could ever have started if the first games took 12 hours to play.

While designers might concede in principle that a playing time of a single evening or afternoon is the ideal, in practice wargame playing time tends to run afoul of the pursuit of detail. (Detail here is distinct from realism: a battalion-level game is almost by definition more detailed than a brigade-level game on the same subject, but it may or may not be more realistic. If you’re reading this, it is likely that you’ve played enough games that combined a high level of detail with a gross lack of realism to appreciate the distinction.) If we look at Stalingrad vs. War in the East for example, the huge increase in the latter game’s playing time is largely the result of the pursuit of additional detail; going from monthly to weekly turns and going from corps to division units. WitE, however, was not a playing-time-is-no-object design. Its air war component, for example, was highly abstracted and imposed a very modest time cost on the game. Partisans were included but their use was discouraged because of the very adverse effect they had on playing time. The 1st edition of WitE had a production component was highly time-consuming, but it was replaced in the 2nd edition with a system that vastly reduced its playing time burden. (The 1st edition system required that units in production be pushed around tracks: every turn every unit on every production track had to be pushed one space forward. In the 2nd edition, a spiral track, in a triumph of mechanical efficiency, completely eliminated the need to move units on the production track; once placed, the player didn’t touch in-production units again until it was time to bring them into play.)

Now, there is a reason that detail is pursued, and that is because it is part of making a game a historical, which is part of making it fun. Historical wargames grow from the soil of history, it is what separates them from abstract games and gives them much of their interest. But if a designer wants to manage playing time, he has to be discerning about the playing time costs of detail. Some detail is expensive in terms of playing time, but other detail is not. Artistic detail in particular tends to be cheap. Non-functional detail on counters or a map, however it may run afoul of the Redmond Simonsen’s austere functionalism, can add detail to a game at little or no cost in playing time. At the other end of the cost scale is detail that dramatically slows down common player actions. (Bloody April is as good a high-cost example as I know of: It took the Terrible Swift Sword game system and brought it to its knees by adding a straggler die roll to every move of every piece.)

Designing for a faster playing time has one other huge advantage in addition to allowing players to actually play your game rather than having it sit on their shelf: it can make your game more fun to play. It is a question of editing for pace. Just as the impact of a movie can be enhanced by removing unneeded scenes, subplots and characters, so too the impact of a game can be enhanced by removing unneeded details, systems, and components. A game is not better for being a mix of good and bad stuff, just by virtue of having more stuff. Better by far to have a game that is all good stuff; leave the bad stuff on the cutting room floor where it belongs.

5 January, 2011

Another entry in my personal history of playing strategic eastern front games. This one is not strictly an eastern front game, since it includes the western allies, but it is my list and I’ll put whatever games on it I want to put on it. Here it is, Battle for Germany:

If there is one thing that Stalingrad and War in the East had in common, it was their aversion to the end of the war, but here is a game about nothing but the end of the war. For that alone it would be noteworthy, but the 2-player set-up where one player takes the western allies and the east-front Germans, while the other takes the west-front Germans and the Soviets, is one of the most inspired design choices I’ve ever seen. As a bonus, the game can also be reasonably played with 3 or 4 players as well. As yet another bonus, the game can be played as a hypothetical red-star-white-star war between the Soviets and the western allies. The game was an instant hit with pretty much everybody and has been through many editions.

One of the other interesting aspects of the game is that it is so much like Stalingrad. It has a small but varied piece mix, a simple move-attack game system, and just uses replacement pieces as its “production” system. The historicity is quite a bit better than Stalingrad, although it makes no pretense of being the “ultimate game” on the subject. (Incidentally, I’ve seen so many “ultimate games” that were unplayable messes that I’ve come to think of the word “ultimate” as meaning “horribly broken” when placed before the word “game”.)

So why did BfG game work?

Well, first, it didn’t make the mistake of WitE (and many other WWII strategic/operational games of WitE’s generation) of having a uniform piece mix. Much of the interest in games with odds-based CRTs comes from the mathematical exercises of looking for the right combinations of magic numbers. (I don’t mean this disparagingly — games that derive their interest from playing with numbers, like Sudoku, are by no means an inferior form of game.) If the numbers are uniform, then much of the fun is drained away. Interestingly, the numbers don’t have to highly varied on both sides; things can actually work fine even if they are varied mainly on one side. In BfG, it is the variety of the German piece mix that principally drives tactical interest. The Soviets or western allies, left to themselves, are a little bland and, at the risk of critiquing a rather nifty classic of a game, it might have added some juice to the game to mix them up a little more.

Second, it has a nice asymmetry to all of its various scenarios and situations. The Germans play nothing like the Soviets or the western allies, and the eastern front plays nothing like the western front. And of course, in the two player game, both players get to play both offense and defense. For a small, simple game, it includes a nice variety of game experiences.

Third, it takes a famous situation, and deals with it in an inherently atmospheric way. The period is the end of WWII, but it is also the start of the Cold War, so having the Soviets and western allies playing as much against each other as against the Germans isn’t just a gimmick to the make the game more interesting, it reflects the reality of the historical situation. Game and history are brought together in a way that is so simple, yet so perfect.

4 January, 2011

Last week I ordered nickel sheet metal to try to make some better playtest pieces for Stavka. It arrived, but I think it is too hard for me to work easily. I will be ordering another, thinner sheet, this time of aluminum. My fear is that the aluminum sheet will prove too soft to hold a shape well (like tin-foil), but we’ll see.

In the meantime I did another playtest with the polyclay pieces. The mechanims of progressively increasing cost for the defense-in-depth pieces (which I’m thinking of renaming strongpoints; in fact, let’s try that out right now: The mechanism of progressively increasing cost for strongpoints...) is working pretty well. Not too sure yet what the magic numbers should be for the prices, but making the price dependent on the number you already have in play makes historical sense (can’t concentrate everywhere) and game sense (the defender has to pick and choose where to be strong), and it plays fast and easily. Looks like a big win.

I’m still not sure on the overall numbers for card draws for each side. One area revamped were the rules for “standing down” armor units. In previous test games, a player could get one additional draw every turn for each armor unit that was stood down. This proved excessive. This time, I tried it as one extra draw on the turn it was first stood down (rather than every turn while it was stood down) and then you had to pay a card to stand it back up again. That may be too little benefit rather than too much, but I don’t have a lot of room here without going into fractional numbers — and I’m not going into fractional numbers. Period. Still thinking about this one.

In this game, the Germans took Moscow in the fall, only to lose it again in the winter. I had the Germans do one more offensive before winter, which in retrospect wasn’t wise. In the spring, the Germans took Leningrad, and then attempted a summer of 42 push on heavily fortified Moscow defenses. The second Moscow offensive proved something of a disaster. I’m not at all sure I’m very good at this game. It seems like I just blunder from one disaster to another. (Because I can see both hands of cards, I try not to take advantage of that to pre-calculate before making my moves; instead I use my gut, and I think I have a stupid gut.)

Anyway, Stavka is coming along. Still lots of problems to be solved, but the design is moving forward at a satisfactory pace.

3 January, 2011

I thought I’d continue with my personal history of playing eastern front games. Now, I don’t intend to list every eastern front game (there were so many, and so many that were so forgettable...) but I did want to hit on the high points: the games that actually had an influence on my thinking about this subject and on game design in general.

Next up is this little number, SPI’s War in the East:

I mentioned in my previous blog entry that the design of Stavka was partly a reaction to Stalingrad, but mine was far from the first game to react to Stalingrad, nor was the direction I’m taking the only direction a reaction to it could take. The history and thinking behind WitE is in part reflected by its early title: Stalingrad II. Clearly, the idea of WitE was that the main thing wrong with Stalingrad was that it didn’t have enough, so WitE had more. Lots more. It was division rather than corps scale, it had four maps instead of one, weekly instead of monthly turns, a mechanized movement phase after combat, Soviet production, extended coverage, including Murmansk geographically and 1944 chronologically (still no 1945 though). It had railheads, fortifications, one-year scenarios for '41, '42, '43 and '44, partisans (even if the design notes recommended against actually using them), it had Sevastopol on the correct side of the Crimean peninsula, it just had lots and lots. It was huge. It was epic.

It was also tactically rather dull. Sure, a division-level game on the eastern front had to work awfully hard to be completely bad; and WitE wasn’t completely bad. Just the scale of the game had a certain captivating quality to it, and as it was my first monster game (and lots of people’s first monster game, I think) I was by no means immune to its spell. I suppose the limitations of the game were actually captured in the design notes that came with it: they recommended that the Russians cut and run at the start of the game to set up a fortified line outside of German supply range and hold that through the end of 1941. After that, well, the design notes didn’t say much, but I can tell you from experience that 1942 was Verdun, the Germans attempting to bleed the Soviets dry through attacks on that line. If they succeeded and punched a hole in it, that was pretty much it. If they didn’t, well, that was pretty much it too. Although WitE included 1944, I can’t imagine that many people played the game past the end of 1942. Personally, 1942 tended to suck the life out of me. Tactics were SO basic: another turn, another set of attacks by 48 or 72 points of German infantry on Russian stacks, with each Russian stack consisting of a 0-3-0 fortification, a 4-4 rifle corps, and a 0-1-10 anti-tank brigade. Turn after turn of the same damn thing. One thing about Avalon Hill’s Stalingrad, the Soviet line always consisted of a mix of units, some strong, others weak, and some in strong terrain, the Soviets were always thinking about where to set up next and the Germans were always thinking about where the best attack opportunities were and what was the best combination of units to do it with. And oddly enough, even though WitE was the one with the mechanized movement phase that was supposed to add mobile warfare to the game, Stalingrad, slow as it was, was actually more mobile than WitE — I didn’t really count WitE’s recommended 1941 Soviet cut-and-run as mobile warfare. I never had a slower-going experience in wargaming than those 1942 German attacks on that Soviet fortified line.

So where did WitE go wrong? Well, in one sense it didn’t. Even if you never finished a game, if you looked at time spent vs. dollars spent, it was a pretty good deal. But it really was the anti-Stalingrad in one sense: while Stalingrad was a game that you could play over and over again and still find it new, WitE was a game that you couldn’t even play once without finding it old. And I did find it old, but for a while at least, WitE did exert a spell, even if it didn’t last. And it is a game I don’t have to struggle to remember. That damn Soviet 0-3-0,4-4,0-1-10 stack is still tattooed on my brain, decades later.

2 January, 2011

Happy New Year everyone. Playtesting on the 2nd edition of Bonaparte at Marengo continues. The testers have updates on the rules and map that include the new “sudden death” marginal victory for the Austrians. The last playtest was actually a good test in that it exposed a weakness in the game, enabling it to be corrected prior to publication, which is, after all, what playtesting is all about.

I am (still) working my way through The Wages of Destruction, the economic history of Nazi Germany I’ve been reading for Stavka and have only now finally, at over 400 pages in, reached Operation Barbarossa. One of the interesting things is the discussion of why, although Germany in theory had the resources of all of non-Soviet Europe to draw upon, which both in population and pre-war productive was actually on a par even with the United States, so little was made of it. Essentially, the problem was the blockade. The portions of Europe the Germans occupied drew critical resources (chiefly oil) from outside of Europe to run their economies and once the areas fell under German control, they also fell under the blockade and all these sources were cut off. The result was a huge contraction in their post-occupation economic capability, such that they could contribute very little to the German economy. Their main contribution was not ongoing production, but one-time resource contributions from being looted of pre-war inventories of oil, copper, grain, production and transportation equipment, etc. Interesting.

Speaking of Stavka, I’ve wanted to talk a little about my personal history playing strategic eastern front games. Like most gamers of my generation, I started with this guy:

Now, it is very far from my purpose to beat up on a 50-year old wargame, especially the one that introduced me to the subject and which was as solid a gamer’s game as Avalon Hill’s Stalingrad. Oddly enough, one of its weaknesses as a simulation was also one of its strengths as a game: it never had many Soviet pieces in play. However, the pieces it did have were varied enough to keep the Soviet army from being boring, and those river lines and cities (not to mention the dubious mountain range south of Leningrad) made for surprsingly subtle terrain to attack and defend. The result was a game that had enough variety to be interesting, but not so much as to be overwhelming: Stalingrad invited efforts to master it while simultaneously frustrating those efforts.

Now, in terms of game play, what I’m trying to do in Stavka has nothing really to do with Stalingrad, and in fact is in many ways a reaction against it. (A game that can still inspire any reaction at all decades after publication must have been doing something right.) Stalingrad was essentially a slow, attritional game that only covered the part of the war where the Germans were on the offensive, whereas what I am trying for with Stavka is speed and armored penetrations to keep the enemy off-balance, while covering the entire campaign from Barbarossa to Berlin. Still, there was one area where Stalingrad succeeded where I will be very pleased to do half so well: to captivate the people who play it enough that they will want to keep coming back for more. Whatever else I have wanted to do with my games, I have always wanted to do that, and this crudely produced, deeply ahistoric game is still a benchmark as to what success in that endeavor looks like, almost half a century after it was made.

Anyway, with the end of the holidays I plan to resume Stavka testing. Lots of issues there that still need to be resolved.