Home Store Products Research Design Strategy Support News

27 March 2010

 Products The Guns of Gettysburg Design Diary Rules are Rules

Rules are Rules

In the last diary entry I attached a copy of the rules and invited feedback. I got some (Thanks to everyone who responded!) and have been making edits in response. Now, I didn’t follow every suggestion, which wouldn’t even be possible since often different people offered contradictory suggestions, but I read everything I was sent, and thought about all of it — at least I tried to, but of course it is always possible to miss something. Anyway, below I’ve posted an update showing the rules as they currently stand.

Gettysburg Rules

Click on the image above to open in its own window

Some of the comments of course were just typo corrections of various sorts. (Always welcome!) These were easy corrections to make. Other feedback regarded wording, either complaints about lack of clarity, excessive verbosity, or suggestions for improvements. These were welcome too. Other feedback was in the form of questions: Now questions sometimes reflect a fault in the rules, and sometimes not. Everyone misses a rule now and then, and some questions were no more than that. Other questions, however, reflected problems in the rules themselves. One recurring theme was that certain key ideas were not made as clear as they should have been.

The basic concepts of area, position, positions bounding areas were all beefed up in this version of the rules. Some of the fault was mine (I never defined exactly what it meant for a position to "bound" an area, for example), and part of it was just NT-itis: people carrying into the game preconceptions about how it worked based on NT (most commonly) and not processing rules that did not fit those preconceptions. Part of the problem for readers, of course, is that game rules tend to be best understood when you try to actually make the game go, and the readers don’t have the game and so can’t do that.

In addition to content issues, there were also some formatting problems that were addressed: Most of the rules are formatted by style sheets, but in some cases there were style sheet overrides in place that should have been, causing minor layout problems here and there. These have (hopefully) been cleaned up. Another change was actually I decided on a global font change. Now, for basic text I have long used Times New Roman as my default font. It’s a good basic font that doesn’t call attention to itself, allowing the reader to read the text without even seeing the font, which for something like rules is what you want. However, I decided that it was high time for a "font-off" and I compared the rules in a variety of serif fonts. The winner was Minion, which is a little lighter and more open than Times, and I think it makes the rules look a little less heavy and more open. To be honest, the difference is prett subtle, and if I hadn’t mentioned it, probably few of you would have noticed (some may find it hard to see the difference even after being told). The effect isn’t really intended to be consciously visible: Minion, like Times, is a font that doesn’t really call attention to itself, so the hoped-for effects will be mostly subconscious.

General Structure

Quite apart from discussing edits, I want to talk a little about the problem of structuring a rule book. A rule book has basically two functions: to teach the reader how to play the game and to serve as a reference when the reader has questions that come up during play. The first requires that the rules have a narrative flow, and carry the reader forward without forcing him to jump around in the text. The second requires that the rule book have a structure that helps the reader find a particular rule. Thus, a good rule book will have something of the novel and something of the dictionary to it, without really being either.


From a narrative point of view, there are two basic approaches: the first is top-down: you start with the general concepts of the game and then add details as you go along. The second is bottom-up: you start with the details of the game and build up to the general concepts. A top-down approach will start by reading like a novel and end by reading like a dictionary, while a bottom-up approach will go the opposite way: it will start by reading like a dictionary and end by reading like a novel.

Now I’m not going to tell you that one approach is the bad approach and the other is the good approach. I will tell you that I find the top-down approach very difficult to write. I want to express myself very clearly, but when I write top-down, I can’t do that because I haven’t yet defined the precise meanings of terms I want to use. And so I more or less default to bottom-up because I feel comfortable doing it. If I ask myself whether (from a reader’s point of view) my efforts have been successful, I would say only partly. My rules have been largely complete and the answers to people’s questions are generally there and unambiguously expressed. That's the good part. The bad part is that people often find my rules difficult to understand conceptually. While there is no doubt that the lack of any significant common ground with other wargames is partly the cause, I have no doubt that it is also in part the result of the fact that I write in a bottom-up style: the reader is forced to learn a lot of terms without really understanding why he is learning them, and the conceptual framework (to the extent that it comes at all) only comes later.

Now you might ask, if I’m aware of the shortcomings of my approach, why don’t I fix it? Well, it isn’t as easy as that. The top-down approach has problems of its own in that it tends to produce ambiguity; the higher level rules that precede the lower level rules necessarily lack clarity (the definitions and terms needed for precision have not yet been presented) and this can easily lead to ambiguous, even contradictory, rules. The intent in such cases is that the lower-level, more detailed rules have precedence, but having rules that are not consistent is a dangerous thing to do, especially when the rules are being used not as narrative, but as reference, where the reader, rather than taking in the whole context, dives right in looking for a particular rule. I’m not saying the top-down technique can’t be done well, but I can’t say that my own attempts at it have been really satisfactory to me. Anyway, speaking of the rules as reference, I’m going to close this topic and move on the subject of structure.


The basic structure of a rulebook is a hierarchy. The first question in developing a rule book hierarchy is how to build it. In general, I have found that there are two common approaches: one is to break the game down into objects and the other is to break the game down into procedures. What I mean by these is perhaps most readily explained by example. Here are two possible top-level structures for the rules for The Guns of Gettysburg; one by object and the other by procedure:

Rules By Object
1. Game Components
2. The Map
3. Tokens
4. Time and Reinforcement Tokens
5. Artillery Tokens
6. Field Works Tokens
7. Objectives
8. Blocks
Rules By Procedure
1. Set-up
2. Sequence of Play
3. Turn Duration Phase
4. Drawing Battle Tokens
5. Withdrawal Moves
6. Attack Moves
7. Reinforcement Moves
8. March Moves
9. General Order Declarations
10. Ending the Game

The guiding principal of the object structure is that the game is composed of things, and every thing has attributes; every thing also has parts and each part has attributes (and can in turn be composed of still smaller parts) and so on. When structured by object, the rules consist of a systematic breakdown of these objects, their attributes, and their parts. On the other hand, the guiding principal of the procedural structure is that game play consists of activities, and every activity has a sequence and sub-activities within that sequence, and things that are allowed and disallowed in each activity and sub-activity.

Now, each method does some things inherently well and other things inherently badly. If, for example, you want to know all the rules concerning obstructed terrain, the object structure makes it easy to find: you know that in the map section there will be sub-sections for each type of terrain, and for each type of terrain will be all the rules concerning that type of terrain. On the other hand, if you want to know all the rules for obstructed terrain in the procedure structure, you have your work cut out for you: there might be obstructed terrain rules in any of the sections and there is no inherent and obvious way to find them all.

No matter which method you use, it is possible for the reader to want to know something that cuts across the grain of your structure, so there isn’t really any “right” answer. In my own rules, I primarily use a procedural structure, but I do include some object elements as well. Here, for example, are the rules for GoG, showing you which section generally fits in which category (not all sections are 100% one or the other, but this gives you a good general idea):

Object Sections
1. Introduction
2. The Blocks
3. The Map

9. Fields of Fire

18. Cavalry
19. The Iron Brigade
20. Winning the Game
21. Full-Obstructed Positions (Advanced)

23. Historical Reinforcements (Optional)

25. Team Play (Optional)
Procedure Sections

4. Tokens and Set up
5. Sequence of Play
6. General Orders Declaration
7. Turn Duration Phase
8. Drawing Battle Tokens

10. Block Movement
11. Reinforcements
12. March Moves
13. Attack Moves
14. Withdrawal Moves
15. When a Block Can Move Twice
16. Objective Phase
17. Night

22. Extended Bombardment (Advanced)

24. Game Balance (Optional)

The procedural bias is actually somewhat stronger than it appears here, because all of the long rules sections are procedural. By length, probably about 75% of the rules are in procedural sections, with only about 25% in the object sections. I have two general motives for using object sections: one is to introduce terms that are not procedure-specific (the blocks and map sections specifically serve this purpose) and also to extract exceptions and digressions from the procedure sections, to make them clearer and tighter. An example of this is the rules for reducing blocks for losses. This is a about a half column of text. Originally it was in the rules for attacking, which is where the reader first needs to know it, but in that location it was a large digression that completely derailed the description of the attack procedure, which is the rulebook’s longest and most complex procedure, and where consequently any digression was least wanted. And so I moved it out of that section and moved it into the section on blocks. Is that an ideal place for it? Frankly, no, but it isn’t in the way there as there isn’t otherwise much going on in that section.

Of course, as I mentioned earlier no matter what organization you use, the reader will still sometimes have questions that your organization doesn’t lend itself to answering. One approach (one I don’t personally like) to solving that problem is redundancy: you include exactly the same material in multiple places in the rule book. For example, the rules for the effects of obstructed terrain on movement might be duplicated in the rule book: once in a terrain section and again in a movement section. This sort of redundancy ensures that whether the reader looks in the rules for terrain or whether he looks in the rules for movement that he will find the rules for the effects of obstructed terrain on movement. Like most tools, I wouldn’t say that redundancy is something I would never do, only that I try to use it sparingly. One example of where I do use it is in the close combat modifiers for the Iron Brigade: they are listed once in the section on the Iron Brigade, and again in the list of close combat modifiers. Redundancy, however, if extensively used, can be quite disruptive to the narrative function of rules: the length of the rules is increased, the time to read them is increased, and the reader gets recurring feelings of déjà vu as he reads them.

A less intrusive solution is, simply, an index. I’ve included on in the GoG rules. (I should have done this for Napoleon’s Triumph: it needed it.) An index isn’t ideal, since the reader still has to flip around in the rulebook to find the information he’s looking for, since the text isn’t all gathered in one place, but it is better than nothing, and unlike redundancy, it doesn’t impair the narrative experience.

There is more that can be said on the subject of rules writing (a lot more to tell the truth), but I’m going to leave off the subject for now. Before I sign off though, I have one small, well, not that small, unrelated subject to discuss:

A Dash of Color

As anyone who has read much of what I’ve written knows, I have a strong interest in the graphical aspect of game design and color in particular. I seldom, however, directly discuss games other than my own. However, because I think the subject interesting, this time I’m going to. My game, of course, is very far from the first Gettysburg game ever made. I think Richard Berg must hold the record for Gettysburg games; I’m not sure how many times he’s dealt with the subject, but he’s treated the subject at a regimental level five times: the two editions of Terrible Swift Sword and the three editions of The Three Days of Gettysburg. Now I admire Berg's work in some ways, and have reservations in others, but it isn’t actually his work that I’m going to discuss here: it is the work of his games’ graphic designers.

The 1st edition of Terrible Swift Sword had its graphics done by Redmond Simonsen, who has certainly had a greater influence of wargame graphical design than any other man. His chapter in the old SPI book Wargame Design: Graphical and Physical Systems Design, has been rendered obsolete in many of its recommendations on tools and techniques by the advent of computer graphics, but his thoughts and views on the concepts of design are still relevant today. A picture of his work in Terrible Swift Sword (taken, like all the pictures I’m using, from the site Board Game Geek, but hopefully within the limits of fair use) is shown below:

Terrible Swift Sword 1st Edition

The above image is very reflective of Simonsen’s work for SPI in that period. It is graphically quite simple by modern standards, but its simplicity is not solely the result of the technical limits of the period: it is also the result of Simonsen’s ideas as to what is and is not good design. The area shown is the southern end of the battlefield, with Warfield Ridge in the foreground and Round Top in the upper-right corner. The focus in the picture is too soft for the pieces to be readable, but it looks like it might be the historical attack by Hood’s and McLaws’ divisions on the Union left flank on 2 July.

One graphical decision that is very Simonsen is the representation of The Peach Orchard, which is represented in the game by the words “The Peach Orchard” under the Union blocks on the left side of the image. In spite of the label, no actual orchard is shown. As to why, it is partly because TSS had no rules for orchards, making it unnecessary from a functional point of view to represent them on the map. This left their representation as a purely graphical question: they could be excluded or included at Simonsen’s whim. As to why they aren’t there, we can get the answer straight from Simonsen in the SPI book:

“The term ‘decoration’ is used to indicate those graphical elements which have no practical bearing on the utility of the components ... Decoration is information – unnecessary information — which if present in overabundance distracts the player from the truly important, game-play information he must have. ... there are some elements of decoration that I am dogmatically opposed to. First on my list of such elements is the placement on maps of extensive terrain that has no effect on play whatsoever. There’s nothing sillier than (for example) a large swath of desert glaring at the player when that desert is no different from ordinary clear terrain. If the [graphical] designer wishes to impart the fact that there’s a desert on the map, he could much more reasonably place a simple line of type indicating the name of the desert.”

And so we have a peach orchard indicated by name but without any trees. Now it is easy to think this a little odd, and I think it is a little odd, but let’s back up a minute and look at the entire visual composition of Simonsen’s game: He uses four ink colors for four different types of terrain (green for woods, grey for slopes, orange-brown for ridges, and blue for streams) with the result that it is very easy to pick out the different terrain types; the map really does the work for you. It does use decoration, but in subtle and unobtrusive ways: the tree-top detailing of the woods and the frayed edges of the rides and slopes are decorations that do not introduce any visual confusion, and in fact subtly help us identify what we are looking at. Most of the map, by the way, is the fifth “color”: the color of the paper, which is used to indicate clear terrain. (Simonsen was a great believer in using uninked paper as the default color for map.) If we look at Simonsen’s work in this way, we can see the advantages of his design approach. The functional aspects of the map are brought out about as strongly as it possible to do, while the color of the map as a whole is restful and easy on the eye.

Now there is one further thing I want to talk about, and that’s the counter design. The TSS counters are very basic by modern standards, but are clean and uncluttered. The use of color is very limited: only four background colors are used (blue and light blue for Union combat units and commanders respectively, and gray and red for Confederate combat units and commanders respectively) with only a touch of accent colors (only one accent color per counter). The very simplicity of the use of color in the counter design, however, strengthens the overall composition of the game in play. In looking at this picture, we can see at a glance what the situation is: without knowing, we can see that we are looking at two armies and can tell which is which. The only thing I would characterize as a mis-step is that the red used for Confederate commanders is really too strong. It is the only really bright color on the map and it draws our attention to it excessively. Even if we think our attention should be directed so strongly to the leaders, we are left unable to explain why then, the Union leaders are an inconspicuous light blue. For the Confederate leaders to be so visually strong while the Union leaders are so visually weak, is (I think) just a flaw in the visual design. Still, there is one other thing worth calling about about this, the colors used in the units are never used for large areas of the map, which gives the units a nice visual pop (the thin gray lines for slopes and the thin blue line for the stream do not confuse the eye by looking like blue Union units or gray Confederate ones). The Confedeerate gray, is of course, weaker than the Union blue, and it is easier to see the Union units as a consequence, but overall I would say that the color space for the game as a whole (map and counters together) is very well-managed and one that even modern games, despite the technical advances in color printing, do not often do as well.

Having dealt with the 1st edition of TSS, let’s move on the the TSS’s 2nd edition:

Terrible Swift Sword 2nd Edition

I’m not actually sure who did the graphic design for the 2nd edition, which was published by TSR. A large functional change — the use of elevation coloring to indicate relief instead of the slopes and ridges in the 1st edition — had a huge impact on the graphic design and is by far the most obvious and striking difference between the 1st and 2nd editions. It is not at all to my purpose here to discuss whether from a game play point of view this was a good or bad thing, but from a graphical perspective the impact on the whole was bad. As much as I wish it were otherwise, I just don’t have much good to say about this game in terms of its graphics.

The terrain shown in this image is generally the same as the previous image, but with the perspective reversed. Before we were looking from behind the Confederate lines, this time from behind the Union lines. We can at least see that the orchards now have trees, and in spite of Simonsen’s views on such things, I can’t help but see that as a generally good thing. (I don’t remember whether the 2nd edition had rules on orchards, although it probably did, making this a functional and not a decorative change.) We can also see the boulder fields of Devil’s Den, which the Simonsen map also lacks (for the same reason that Simonsen’s orchards lack trees, we can be sure — a conscious policy of excluding terrain that does not affect play).

Still, in spite of the expenditure of a great deal more ink than the 1st edition, the overall effect of the 2nd edition is (I think) visually inferior. The first problem is just that there is a huge drop in graphical sophistication. Everything is more crudely drawn and less professional looking. If we look at the woods, for example, the detailing that marked out tree-tops in the Simonsen map is wholly absent here: what we have instead is a splotchy green, which in addition to not being very attractive, is also visually far too close to green used for elevation; the eye wants to read the woods as being an elevation color, and it takes real effort not to. Finally, the map includes some color choices that are just really bad choices. The pinkish shade used for elevation is, I think, very hard to justify, and the brighter pink used for buildings even harder.

If we turn away from the map to the counters, we also see questionable choices. From a functional point of view, there is a real improvement in the use of colored bands on the units to indicate organizational affiliation, which does make it easier to figure out which units belonged to which organizations, which undoubtedly is a time-saver during set-up and play. There is even a graphical benefit in that being able to see this can make improve the visual experience and make the game easier to read. A related change is the use of two different colors for higher vs. lower ranked commanders, again with the idea of making identification easier. Because there are functional gains, these changes are respectable visual design choices. Still, there is a visual cost to the colored bands in that they break up the counters as a color mass and diminish the visual impact of the pieces. If you compare the photos of the 1st and 2nd edition, you can see that the Union pieces are visually stronger in the 1st edition than the second. Of course, it is not only the color bands that are responsible: the stronger colors of the background also play a role; with so much going on in terms of map color the counters have a lot more to compete with. The Confederate army suffers from this even more than the Union: light grey is a very low-contrast color (it is in fact an excellent choice for something you want to have blend in) and particularly against the pink background, the Confederate counters all but disappear. Also, for the Confederates, the presence of two colors for commanders (red and violet), neither of which is remotely related to the gray used for the combat units, breaks them up visually; when you look at the Confederate side of the board, your eye immediately goes to the red command counters, then the violet, and then has to hunt to find the combat units. The result is a weaker overall presentation of the game than in the 1st edition, in spite of the fact that the 2nd edition contains more map information and was (no doubt) more expensively produced.

The 1st edition of Terrible Swift Sword was published in 1976; I’m not quite sure of the date for TSS's 2nd edition, but think it was probably in the early eighties. The three editions of The Three Days of Gettysburg were published much later: in 1995, 2000, and 2004. I don’t have any images in the first two editions that would fit nicely in this set (unfortunately), but here is an image from the 3rd edition, again showing the area from the Union left flank on 2 July:

Terrible Swift Sword 2nd Edition

The art for TDOG (3rd edition) is credited to Mark Simonitch (map) and Roger McGowan (counters). Functionally, the Simonitch map meets much the same requirements as the TSS 2nd edition map: it shows elevation, orchards and woods, main roads and local roads, boulder fields, swamps, and farm houses. While there are some functional differences, the main differences between the two maps are in the quality of execution. Part of this is because of the improvement in tools in the interim (computer graphics have vastly simplified adding fine detail to maps) and partly due to much better graphical taste and judgment. While the TSS 2nd edition map is just a bit of a mess, the Simonitch map is a thoroughly professional product. I do not know if Simonitch shares Simonsen's views regarding the use of decoration, but the use of decoration is as restrained here as if he did.

The counters in TDOG carry forward the use of color for command identification that were present in the TSS 2nd edition, but they are improved in some important ways. A touch of secondary color is used to identify brigades with the bands being left to identify divisions. In the counters, as well as the map, the technical improvements made possible by computer graphics are very much in evidence: the small symbols are much more detailed and are in multiple colors. As a result, while the counters from the older games tend not to hold up well under close examination, the modern counters do. Another counter-design difference we can see is the change from gray to butternut for the Confederate combat units and from red to gray for the Confederate commanders. Taken in and of itself, the counter design is open to no serious criticism. The nearest comment I can make to a criticism is it lacks the flair of Simonsen's asymetric design (symbol on the left, strength on the right) with its more conventional centered symbol.

It is only when the counters hit the map that we can find real fault. While the Union units stand out reasonably well, the Confederate counters are far too close in color to the light brown used for elevation. This is less a bad design decision as such than the result of a very crowded color space. The need to include so many shades of elevation color drives the map art to use browns and greens for elevation shades, which leads to a collision with the Confederate counter color. Ironically, the gray Confederate commanders show up better than the butternut combat units; while gray may usually be a problematic low-contrast color for counters, here I think would have worked better than the butternut.

For the Confederates at least, the colored bands work well here, in part because they’ve been moved from the middle of the counter where they break up its shape, to the top where they don’t. The visual potential to help rather than hurt the overall look of the game is realized here I think: at the division level, there are enough units with each color that they form a recognizable mass to the eye, rather than a riot of unmatched colors: we can easily see Hood’s division to the right with its yellow bands, and McLaws' division to the left with its green bands, while both Hood and McLaws are both easily recognizable as Confederate. The secondary colors used for brigade identification are small and quiet: we can see them when we want to, but they don’t distract from the overall effect. Well done.

There is another fault in the visual effect of the TDOG map: the markers. I must point out though that both editions of TSS had the same problem, even though the illustrations I use don’t show them, and I would not at all want you to think that they are a problem introduced by McGowan: in fact, his markers are probably the best of the three. The problem is that they still aren’t very good. If we look at the Union side, we see that the center of the Union line is obliterated by white breastwork markers with heavy black/brown markings. The white color itself is a visual shock, not at all helped by the fact that they hide everything about the units underneath them. Given these marker’s sparse information content I would think that surely something could have been done. Perhaps producing an extra set (there aren’t that many in the game), one against a blue background and the other against a butternut background; perhaps a reduction in size, so that rather than being 1/2" square, they were 1/2" x 1/4", thereby allowing information about the unit beneath them to be exposed; perhaps a change of positioning so that instead of putting them in the hexes, they went in the (usually) empty hexes in front of them; leaving the units literally behind the breastworks instead of under them.


It is beyond me, I think, to put together anything like a unified conclusion to a diary entry so utterly split between two such disparate subjects. However, I don’t think the first subject is actually much in need of one, while the second subject is. As to why I talked about the graphical design of those three games, I admit it grew more from a deep interest in the subject matter than from any particular application to The Guns of Gettysburg. All of the Gettysburg games whose graphics were discussed were of the same scale (ground and unit), all were hex-and-counter based, and all were united by being products of the evolution of a single game system. The changes in technology over the history of these games have made possible great improvements in physical presentation; however, I think that there are two process splits that, if healed, could make possible even more great improvements. The first of these is the split between map and counter design. Simonitch’s map is really quite good, as are McGowan’s counters, but the two together are less than the sum of the parts, Simonsen's map and counters are inferior to both, yet his game as a whole is more harmonized. I would argue strongly that colors used for the map should not be used for the counters; that a good game in play should display the armies strongly against the map. (If Simonsen had a dogma about excluding non-functional terrain from his maps, I admit I have a dogma about this subject.) The second, and more important, split is between game design and graphic design. Rather than treating graphic physical design as an afterthought that happens after the game designer has finished his work, the game can be improved by introducing it much earlier in the process, and critically feeding back and modifying the game design with the aim of actually making it more attractive to look at and easier to play. Even if the person doing the graphic design is not the same as the person doing the game design, as there often is, that doesn’t mean that communication can’t be a two-way street: graphical design (I think) can and should push back on game design with the end of producing a more attractive and better product. Graphical and physical design isn’t a limit on good game design: it is a weapon to create good game design. It is only the habit of designing the game first and the graphics and physical components afterwards that keeps the potential of physical design from being realized.