Home Store Products Research Design Strategy Support News

27 July 2009

 Products The Guns of Gettysburg Design Diary Gettysburg Ratatouille

Gettysburg Ratatouille

I had two reasons for doing The Guns of Gettysburg: (1) I had some ideas about fields of fire and wanted to try them out, and (2) I wanted to do a game on Gettysburg. The sad fact is, however, that there is rather a gap between the first and the second: To make a Gettysburg game, there were a lot of design problems other than those about fields of fire, and for those I had no particular ideas. This diary entry will review a list of some of the open design issues that came up during the design process, and what happened with them.

Artillery Organization

I can’t say that I was ever really satisfied with the way artillery was handled in my Napoleonic games. Napoleonic artillery was actually more or less dispersed through the organization of the army, but my games represented it as concentrated in a small number of pieces. (In the case of Bonaparte at Marengo exactly one piece per side.) Now, this wasn’t all bad, because armies could and did concentrate their artillery on occasion to form grand batteries, which my artillery piece design could reasonably represent. In fact, at Marengo, both the Austrians and the French did form grand batteries. Artillery, however, wasn’t always used as grand batteries, but I was kind of locking the players into that usage model by the design of the pieces.

Now I did think about using this same method for Gettysburg that I had used in the other games, but I didn’t really want to. Certainly if I couldn’t think of anything I liked better, I could do it that way, but I was very much open to alternative approaches, although when I started the game I had nothing particular in mind as to how to do it. All I knew was that a more flexible system was desirable if I could come up with one.

Artillery Ammunition

Closely related to the organizational modeling of artillery was whether (and how) to model ammunition supply. Gettysburg was a long battle, and the Confederates in particular had not brought much ammunition with them. (The absence of any Confederate re-supply system for their army in the campaign is the reason that some historians have characterized the operation as a large-scale raid rather than an invasion per se.) Historically, in fact, the Confederates pretty much used up their artillery ammunition supply during the fighting at Gettysburg, and could not have fought another battle even apart from the losses in men they had suffered. While I was strongly opposed to any sort of heavy bookkeeping load for the game, including the problem of ammunition supply would certainly add to the game if I could come up with an easy way to model it.


Another open question was whether to do anything about leadership in the game, by which I mean modeling individual leaders, which neither the Marengo game (which had no leaders) nor the Austerlitz game (Murat was treated no differently than Bagration) did. Should I do something with this at Gettysburg? Now I can’t say that I regretted not doing this in the Napoleonic games. It is, of course, very common for games to rate commanders, and there are even some games where I think it is well-used and important to the game. One positive example would be Frank Davis’s Frederick the Great, where it was really central to the way the whole game worked. However, I have seen many more games where it just extraneous chrome complicating the game without really changing the way it played. (Or worse, changing the way it is played in bad ways that reward gamey and anti-historical tactics.) I thought that if I had a good idea where it would really add to the game and make a better game in a substantial way, then I would go ahead here, but that I would otherwise just pass.

Leader Casualties

If Gettysburg was anything, it was a dangerous place for senior commanders. Of the nine division commanders in Lee’s army, three were seriously wounded or killed: Heth, Pender, and Hood, and two of their replacements (Pettigrew and Trimble) were casualties as well. On the Union side, of the seven corps commanders, three were casualties: Reynolds, Hancock, and Sickles. There was nothing remotely like this in the Napoleonic battles I had done; should it be included in the Gettysburg game?


Friction is the term Clausewitz used to describe the myriad small difficulties that come up during war that make everything work more slowly, less efficiently, or not at all. For example, orders sometimes get lost, or misinterpreted, or delayed; sometimes subordinates have to act without any orders at all, based on the situation as they see it. Sometimes units get lost or delayed; bridges can wash out, maps can be inaccurate. In short, lots of things can happen that can prevent a plan from being executed as its author expects even without any action by the enemy. Now, the limits on the number of moves that could be made in my two Napoleonic games were basically simulations of friction. The Austerlitz game was the more sophisticated of the two, in that it really captured the breakdown of the armies as the battle developed, which the Marengo game did not attempt.


Closely related to friction is uncertainty: limits on the player’s knowledge of the current situation and what might happen in the future. Traditional hex and counter wargames use combat result die rolls as their major, often sole, engine for uncertainty in the design. In my Napoleonic games, limited intelligence, concealing the type and strength of enemy units, was used for much the same purpose. However, limited intelligence of this sort only really works as long as their is variability in the types and strengths: if all the units are of the same type and same strength, it doesn’t accomplish anything. In the Napoleonic battles, the various grades of infantry and cavalry (line, light, elite, guard, and so forth) together with the mix of arms (infantry, cavalry, and artillery) had worked pretty well for this purpose. However, in looking at Gettysburg, I was looking at basically infantry armies, particularly if I didn’t represent artillery with blocks, and infantry armies that were not formed into different grades as Napoleonic armies were: no grenadiers, no jaegers, no guard, etc. While I did have a plan to differentiate between different quality levels among the infantry units, it was going to be in a substantially narrower range than it had been in the Napoleonic games, and the result was going to be a radical reduction in the amount of uncertainty in the game coming from limited intelligence regarding the blocks.

It was this problem that really forced me to do something and to take a look at all of the previously discussed design issues from the perspective of whether they could be used to raise the level of uncertainty in the game to what the game needed to have.

Chop Them All Up, Mix Them All Together, And Voila! You Have Way Too Much.

Well, with the above list in mind, I got the idea that I could use cards and get all of those things into the game, and pick up a fun strategy element of card management in the bargain. Full of enthusiasm, I set to work. Now as it happens, I had never put cards in a game before, and wasn’t sure about the best way to make prototype cards for playtesting. Fortunately, there are people who will sell you blank, micro-perforated card stock that you can print on, and a coating to apply to the cards so that they can be shuffled and dealt more or less like professionally purchased cards. All you need to do is supply the artwork, which I set out to do.

Card Backs

You can see the design for the card backs to the left. I knew enough about cards to know that you should keep the back of the cards fairly busy: it helps to hide marks and keep any of the printing on the other side from showing through. Now the cards were one of those ideas that just didn’t work, but I still love that graphic design I did for the back of the cards. I almost would have kept cards in the game just so that I could have it. The background graphic is the map of the battlefield, stripped of all its regulatory elements, and the foreground is a grid of all the state flag designs used in the game, labeled with the names of the states.

I never actually had rules for the card play, but the basic idea was that you would draw a card a turn (or something like that) and would have some maximum hand size, like 7 cards or something. Different cards would be good for different things, and so you would manage the contents of your hand according to what you wanted to do.

Card Faces

So why didn’t it work? Well, for starters look to the right at what one of the card faces looked like: Yikes! There is just way, way too much going on here. The cannon symbols in the top-left and bottom-right corners were artillery bonuses; for example, this card could be played to give an artillery bonus to McLaws (if played by the Conferates) or Sickles (if played by the Union). It could also be played to give a leadership bonus to McLaws (if played by the Confederates) or Sickles (if played by the Union) or it could be played to kill/wound Sickles (if played by the Confederates) or McLaws (if played by the Union), or it could be played for some sort of bonus at 6:00PM, or it could be played for some kind of bonus in the north-west corner of the game board.

Any sober, sensible person looking at this would think, right off the bat, that at the very least some serious editing was needed here, and they would be right. I, however was not a sober, sensible person. I was greatly smitten with the design I had done for the backs of the cards (Just look at it — doesn’t it look great?), and was not in anything resembling a critical mood. And so I plowed ahead. I traced pictures for the little clockfaces and the cannons. I scoured the Intertubes for high-quality photos of the generals for both sides, and trimmed, adjusted, and tinted them. I scoured my clip-art collection for the little banners around the text. I carefully went over the map, dividing it into regions and created the little pieces of artwork for each region. I laid out a spreadsheet detailing what was supposed to go on each card to generate the combinations. I tried layout after layout, tweaking and adjusting each element so that it was just right. Then came the big moment: I printed out my first test deck, sprayed on the coating, waited for them to dry, punched them out, put my deck together, shuffled the cards, and dealt my first hands to test out my new game feature.

It was just great. Except that I hated it. The first hand I put together, my mind reeled at even trying to figure out whether I should keep a card or throw it away, or to figure out how I could possibly put together a strategy on how to take advantage of what I had. What the hell had I thought I was doing all that time?

And so I abandoned the idea of using cards and started over.

But don’t the backs of the cards look great?

Time Heals All Wounds

The big crash of my card idea brought the design to a complete halt. For months, nothing really happened with the game and I just did other things. Slowly, although I wasn’t really aware of it, I actually was making progress, mostly in re-conceptualizing the game. What needed to be done? Well, the game really needed an uncertainty engine. That wasn‘t a luxury I could do without, it was a necessity. What else was necessary? Well, the game needed artillery. No artillery = no game. Everything else? Not necessary. Was there anything good about the card idea at all? Well, the resource management idea was a good idea. The implementation was terrible, but the idea was good.

A Fresh Start

Over time reflection led me to the idea of using artillery tokens (die-cut cardboard counters) in a way broadly similar to how I had used cards: players would have a tray of tokens (using a tray like a Scrabble tray) that they would manage like a hand of cards, drawing new tokens from a pool (which was like a deck of cards) and playing or discarding old tokens. The new system would have resource management and it would have artillery, and it would even be able to do artillery ammunition (when you have no more tokens in your pool and your tray, you run out of ammunition) with pretty close to zero added complexity. That seemed like good stuff.

Artillery Token

The design would be kept clean and simple. The number of guns would be the strength of the token, with a range of 1 to 3 guns per token (I would eventually cut that back to 1 or 2 guns per token), and every token would be associated with a particular organization and could only be played in support of blocks belonging to that same organization. The size was .75" per token along the baseline. This would allow two tokens to be played side-by-side in front of a single block. Tokens, however, would be placed on the map only during attacks. The rest of the time each player would keep his tokens concealed in his tray.

You can see the basic tray concept below. My current thinking is to stamp it from a metal sheet. The shape of the tray is well-suited to that purpose. (Partly by its nature, and partly by design.) Battle Tray

If you look at the contents of the tray, you might see some surprises. First of all, the Confederate tokens have yellow stars. These mean the same thing they mean on Confederate blocks: they are hints as to which corps the unit belongs to. This was actually the result of a suggestion by one of the playtesters, who even suggested where to put them on the token. Second, you will notice that some of the Confederate tokens have a corps number on them in place of a division commander name. These are corps artillery tokens and can be played to support any block in the corps. (The Confederates attached artillery at two levels: division and corps.) There is also a token marked •Any•. This token is just a complete wild-card and can be played in support of any Confederate block. Finally, in the middle you will notice a token with a big arrow and what looks like an entrenchment symbol on it. This is the only token in the mix that is not artillery. Instead it is a dual-use token: it can be played during a night turn to build entrenchments on a position. (Something I had always been interested in adding, and which this system handled pretty neatly.) It can also be played during any turn and increase the movement distance of some of your blocks (3 to be exact) by one area for one turn; this just adds a little uncertainty to planning how far enemy blocks can move without introducing much in the way of complication.

Artillery Reserve

The token mix for the two sides is not identical: First, the Union has on-average stronger tokens. (The Confederate tokens are a mix of 1-strength and 2-strength, while the Union tokens, except for their •Any• tokens, are all 2-strength.) Second, the token mix reflects the organizational differences between the armies. The Confederate organization I’ve already alluded to above, but the Union artillery organization was different. They attached about two-thirds of their artillery at the corps level, with one-third held back as the army artillery reserve. The Union army artillery reserve is reflected in the game by Reserve tokens, which are powerful tokens in that they can be played almost as flexibly as •Any• tokens, but are twice as strong.

A Multi-Function Design Element

What I like best about the artillery token design is that it isn’t complicated, but serves multiple game functions. First, it actually models to a reasonable extent the artillery organizations of the opposing armies. Both the Confederate division/corps system and the Union corps/reserve system are modeled. Second, it introduces uncertainty into the game. Since you can’t see what tokens are in your opponent’s tray, you can’t always be sure what will happen in an attack, even if you know the strengths of the blocks involved. Third, it models ammunition supply. Fourth, it provides a way to enforce command integrity: you can only play the tokens for a command if you keep the blocks for that organization together; if you split them up, you can’t play the organization’s artillery tokens. Fifth, it models friction: you might want to make an attack in a particular area, but there is no guarantee you will have tokens that would be useful for that attack: you can wait and hope to draw them, you can attack anyway and hope you get lucky, or you can attack somewhere else based on the tokens you do have.


Ironically, it was because I stepped away from trying to pack so much functionality into the cards and instead go with a less ambitious design that I found I could pick up a lot of the things that I had hoped to do with the cards anyway, as well as a few things (like field works and ammunition) that the cards didn’t do. It is, I think a very successful design element in the game, even if it is a relatively late introduction in the design.

Of course, there is one thing about the artillery tokens that I haven’t shown you: that is how they actually work in an attack. However, this diary entry, like the previous one, is going to end by telling you that an example of an attack is something that will have to wait for another day.