|Products||| The Guns of Gettysburg||| Design Diary||| Quiddity|
In medieval philosophy (My man Thomas Aquinas!) there was a concept called quiddity. The idea of quiddity was that for every thing, there was an essential quality possessed by that thing that made it what it was. In starting a game on Gettysburg, I felt that my first problem was to decide on the quiddity of Gettysburg: What was it about the battle of Gettysburg that made it what it was? It was an important question because that which was central to the battle also had to be central to the game, or it wasn’t a game on the battle. When I worked on Napoleon’s Triumph I had decided early on that the quiddity of Austerlitz was that it was an ambush: if I didn’t have an Allied attack into a French trap, I didn’t have Austerlitz. But what was the quiddity of Gettysburg?
It was pretty early on that I decided that the essential quality of Gettysburg was that it was an accident. It was a battle that neither Lee nor Meade tried to bring about; in fact, they both, in different ways, tried to prevent it. Lee ordered his subordinates not to bring on a general engagement. Meade was preparing an order for his army to assemble for battle miles away at Pipe Creek. Yet the battle happened anyway. Both Lee and Meade arrived on the battlefield to find that it was already in full swing, and both found that stopping it was not a practical option, and both decided that the only thing they could do was to fight it out then and there.
So what does it mean, for a game, for its subject to be an accidental battle? Well, I thought that the most important thing was: “Death to All Plans!” The game would have to subvert and sabotage all attempts at pre-game planning; rather than making their plans before the game started, players would need to improvise, adapt, seize their opportunities, and cope with adversity. That was what the historical battle forced the historical commanders to do, and that was what the game had to force the players to do.
All well and good to say that you will put the players in their historical counterparts’ places, but the sad fact is that the players of the game have a gigantic advantage over their historical commanders: the players know what happened historically. So, if my approach was to work, this advantage had to nullified as much as possible. I decided that my most important tool would be to shake up the reinforcement schedule. Rather than having players know what reinforcements would be arriving when and where, as if the battle was a well-run German train station (Is that an tautology? Are there badly-run German train stations?), arrival times would be randomized, and not just randomized, but randomized and double-blind, and revealed only as the game progressed. With no knowledge of what, when and where units would arrive, players would have no way to pre-plan; they would just have to take whatever the game threw at them and make the best of it.
And so, I was set on my course. But it was not without its problems. It is one thing to have an idea for a game, but it can be a very large job to go from idea to working game.
A minor problem was just the mechanics of double-blind reinforcements: in a computer game this sort of thing was easy enough to do, but in a board game, it was more complicated. I had three things I wanted to hide and randomize: the time, the road, and the unit. After kicking it around for a while, I decided that each army would use two sets of tokens: one set would be shuffled and go face-down on the Time Track (one per hour), and the second set would be shuffled and would go face-down on the various reinforcement entry roads. For each hour of game time, players would check the time track. If their token for the hour was not a dummy token, then players would do the following simple 3 steps to find out where their reinforcements were entering and who they were.
A much more substantial problem was what to do about terrain objectives. With a variable reinforcement schedule, a simple scheme that like that used in Bonaparte at Marengo wouldn’t work. Whichever side got reinforcements faster would have a huge advantage. The structure of the solution was pretty obvious: the objectives had to be tied to the reinforcement rates. My first thought on how to do this was to put more objectives on the map and then require different combinations depending on the rate of reinforcements. A track would be used to indicate the current combination required, and as reinforcements arrived, a marker would move up or down it to record whatever was currently required. I actually implemented this, and it looked like what you see below:
Click on the image above to open in its own window
The objective track is above the map legend, near the bottom-right corner of the map. Fast Confederate reinforcement arrival would result in the marker being shifted toward the bottom, while slow arrival would result in the marker being shifted toward the top. The color stars in the track represented what the Confederates were required to take, so they might need anything from 3 red stars (easy) to 3 blue stars (hard) or anything in-between.
(As a side note, check out the pretty but none-too-functional set-up displays around the board edges in this version of the map, the reinforcement graphics on the entry roads, and the labels in the time track: in this version, each side had a “pool” of possible reinforcements from which they would draw, and at the specified time leaders and blocks — at this time I was thinking the game would include leaders — would be added to the pool; the current system, however, doesn’t work that way.)
Anyway, the above system for handling objectives didn’t last long. It was too coarse-grained, and making it finer-grained meant filling the map with objective stars. And so I decided that a new approach was needed.
Now, as it happened I had a component idea that I’d been playing with for some time in a different context: transparent markers. Rather than print markers on opaque cardboard, print them on transparent thin plastic sheets, like those used in overhead projectors. In that way, the marker could be placed on a game piece to “edit” the piece rather than completing hiding it. (I have a long-standing loathing for markers covering up game pieces.) Anyway, I decided to try out the idea by making movable objectives out of this sort of material.
The basic principal was that the side whose reinforcements were slower to arrive would get to move the objectives as a counter-balance. As a designer, I could tune the game’s balance by changing the number of markers, the rate at which they moved, and their initial starting positions. As a side bonus, I could make the markers come in blue and red colors to differentiate between Union and Confederate-controlled objectives. When an objective was captured, the marker was removed and then replaced by another one of a different color, which was kinda cool.
One thing I learned early on was that I would need a display for tracking reinforcement arrival. This is the odd display in the top of the current board, which you can see below:
(Every time I see that token design, I am reminded afresh of how ugly it is. I can’t believe I never came up with a better design than that. Sigh.)
The way it worked was that as reinforcement tokens were turned up, they would be moved to this display, and the comparative heights of the stacks would indicate who would move the objectives, and how far. If the stacks were the same height, the objectives stayed put. That was sort-of ok, except that there were only seven Union corps and nine Confederate divisions, and the Union corps were bigger than the Confederate divisions. To even things out, I needed some special padding tokens for the Union stack. For this, I made up two special “Received” tokens, one of which you can see to the right. The number on the token indicated when it would be placed on the Union reinforcements received stack; “4th” meant after the 4th real Union reinforcement. Anyway, the seven Union corps plus the two received tokens equalled the nine Confederate divisions, so once all the reinforcements were in, the stacks would be balanced and the objectives would no longer move.
In all of this there is a fairly basic question that I haven’t really addressed (although I think I talked about this in the design diary for Napoleon’s Triumph): Why have terrain objectives at all? What purpose do they serve?
There are three answers to this question. The first is just a basic reality-into-game problem. The game board is just a small window onto the world and has artificial boundaries. In the real world, there was nothing forcing either army to fight on this battlefield. But for a game on the battle, the options to not fight, or to fight somewhere else, must be foreclosed. What’s more, the edges of the game board should be kept out of play as much as possible. (There is a real trade-off here with regard to the area covered by the map and its level of detail: One way to keep the edges out of play is to “zoom-out” the map and have it cover more area, but that means a loss of detail in the central area of interest and large parts of the map would be left almost unused, so in practice you compromise and allow the edges into play to some extent and in exchange you get more detail.) Terrain objectives allow the designer to fix where the action will be with regard to the map frame; both armies have to try to take/hold the objectives or lose the game, and so as the designer you can control where the armies will have to fight by where you put the objectives.
The second reason is simulation-based. Now, in a strategic level game showing entire countries, terrain objectives often can be related to the actual reasons for the war and/or the territory that either side needed to control in order to be able to keep fighting. But in an ACW battle game, like this one, the objectives themselves are not features of any intrinsic importance to either side. Meade’s orders were to defend Washington and Baltimore. Nobody told Meade he needed to defend Gettysburg, still less Cemetery Hill. For battles in this period, the armies didn’t fight for the terrain features because the terrain features were important, the terrain features were important because the armies fought for them. In the Gettysburg campaign, any hill, wood, or town in Pennsylvania or Maryland could be of critical importance, provided only that one army try to hold it and the other try to take it. An offensive campaign was essentially a statement by the army making it that it could go where it would go and take whatever territory it wanted to take, and the defensive counter-campaign was essentially a statement to the contrary: that the enemy army could not go where it would go and take what it wanted to take. A battle was where the opposing claims came to a head: the defensive army plants itself in a position, and the offensive army attempts to take that position. The position itself is incidental: what matters is whether it is held or whether is taken. Coddington, in his otherwise fine book, The Gettysburg Campaign erred, I think, when he wrote that control of the battlefield as a determinant of victory was some sort of medieval anachronism. It was not: it was in fact the final determinant of who won the campaign.
The third reason blends into the first two: it puts the burden of attack squarely on one army or the other. Historically, the burden of attack was on the Confederates; theirs was an offensive campaign, and one without any line of supply. They could not stay concentrated and feed themselves in the North. As long as they were dispersed and on the move, they could live off the land, but once concentrated, they could not do so for long. If the Union army could concentrate in their vicinity, the Confederate army would be forced to do likewise in order to avoid defeat in detail, but they could not sustain such dispositions. So, in order to prevent such an outcome, as a minimal goal they had to drive off the Union army so as to restore their own freedom of movement. (Not to mention, of course, that Lee hoped that a victory on Northern soil would produce political gains for the South that any number of victories on Southern soil had not been able to do.) Thus, the game just from a historical point of view needed to put the burden of attack on the Confederates. Further, from a game point of view, a clear burden of attack makes the game go. When one player knows that to do nothing is to lose, it impels him to act, and his actions in turn force reaction, and thus the whole game is set into motion. Without a clear burden of attack, both sides may become passive and risk-averse, and turn the resulting game into a monumental bore.
Missing in the above discussion is any Napoleonic or Clausewitzian notion of a campaign: to destroy the enemy army and thereby destroy the enemy’s means to resist. This was a deliberate omission. A decisive battle, as a war-ending event, was not a feature of the American Civil War. Even when an army was trapped and forced to surrender itself, as happened at Ft. Donelson and Vicksburg, it was not enough of a blow to end the war. In terms of actual battlefield victories, it is only very late in the war, when the Confederacy was literally being cut to pieces, that you see battles and pursuits that are Napoleonic in their decisiveness: If we look at Nashville, we see that the Hood’s defeat was as total as the Allies at Austerlitz (in military, though not political, terms), and if we look at the retreat of Lee’s army in the Appomatox campaign, we see a retreat as disastrous as that of the Prussians after Jena-Auerstadt (and even more decisive in its political consequences). But these events, I think, tell you more about the state of the Confederacy at the end of the war than about what the opposing Union and Confederate armies were capable of doing to each other in mid-1863. A string of impressive Confederate victories in the year prior to Gettysburg by Lee’s Army of Northern Virginia had never come close to rendering the Army of the Potomac unfit for combat, and even after the defeat of Lee’s army at Gettysburg and the surrender of Pemberton’s army at Vicksburg, the Confederacy was still fully capable of fighting and winning major battles against Union armies.
With regard to Gettysburg in particular, I don’t think that either of the opposing armies was really capable of demoralizing the other in the Napoleonic sense. Lee’s army was punched-out in the historical battle, and incapable of mounting any further offensive operations on that field, but it was very far from demoralized: there was no panicked flight from the battlefield. In fact, the Confederate army remained in the field on 4 July, hoping for a Union attack, while they put their affairs in order and got their baggage trains ready to go back to Virginia. The only prisoners they gave up after the battle were men too severely wounded to be moved, and the Confederates did what they could for them before turning them over to the Union. In terms of hypothetical Confederate victories, the various might-have-beens could indeed have resulted in the Union army being forced to withdraw, but it is hard to see in any of them a victory that would have actually demoralized the Union army. A particular precedent was Second Bull Run, which had relative (though not absolute) army sizes about the same as Gettysburg, and which had a Confederate victory that was about all that could have been hoped for at Gettysburg, but even so the Union army was able to withdraw in good order from the field. Whatever political results a comparable victory might have had for the Confederacy at Gettysburg, I do not think that it could have resulted in the demoralized flight of the Army of the Potomac.
And so, while In my Napoleonic games, there were two ways to win: by taking territorial objectives or by demoralizing the enemy army, I decided that for this game, with the above analysis in mind, there would be no victory by demoralization. Territory alone would decide the game.
While I felt good about the proposed objective system both from a historical and game point of view, it has proven devilishly hard to balance. At the outer limits, the game design had to ensure that a more or less even fight in the area of the map was possible. If Union reinforcements arrived too slowly, for example, it might be that no defense on the battlefield was possible and the early Union arrivals could be driven off the board completely (or just as bad, a battle where the Union army was tucked into a corner of the board). For this reason, the possible range of reinforcement arrival had to be less than completely open. Specifically, I had each player stock the time track with reinforcement tokens once per day, rather than once for the game as a whole; in the initial set-up, for the first 24 hours of the battle, the Confederates would get exactly 7 divisions and the Union exactly 5 corps. However fast or slow one side’s reinforcements started, the other would catch up before the first 24 hours were up. Each side then got the balance of their reinforcements sometime during the second day, but again both sides were guaranteed to be balanced by the end of that day, and then both would remain balanced for the third and final day of the game.
This much of the system was settled fairly early, has always worked well, and I’ve never felt any need to revisit it. However, the initial positions of the objectives, their rates of movement, and the rules governing their movement have been in a constant state of flux as I’ve struggled to find the point at which both sides have a reasonable chance of victory regardless of when the reinforcements arrive. It has by no means always been the case that early reinforcements increased the chance of victory; indeed, in some versions early reinforcement arrival has resulted in objective locations so unfavorable as to make victory almost impossible.
Below are some illustrations of the evolution of the objective starting points. In the first version (3 Stars in Gettysburg), there were 3 stars, and they started in Gettysburg. To make it easier for the Union, I moved the start position to Cemetery Hill (3 Stars on Cemetery Hill). I discovered, however, that having them so far back encouraged very conservative (and boring) Union play, so I moved them forward (3 Stars Forward) but compensated the Union by allowing more rapid Union objective movement, as long as they didn’t let the Confederates capture them. I then decided that 4 stars allowed more strategy, added another one (4 Stars Forward). I haven’t changed the starting positions much since adopting the 4th star, although they have never been completely stable: it seems like I move one or more of them an area forward or an area back between every test.
3 Stars in Gettysburg||
3 Stars on Cemetery Hill
3 Stars Forward||
4 Stars Forward
Click on any of the above images above to open in its own window
There is one feature that was in all the early versions of the system but which was eventually dropped: and that was the abiltiy of the Confederate player, as well as the Union player, to move the objectives. While I quite liked the way that worked in some games, where it introduced a dose of chaos into the proceedings, in others it produced results that were simply bad: in particular cases where a Union reinforcement arrival allowed the Confederate player to move an objective behind his own lines, thereby taking it while remaining in place, could decide the game based on nothing more than chance. In addition to my own reservations, Confederate objective moves were not at all popular with playtesters, and while I generally design to please myself, the fact is that I also always have a little vision in my head when I find myself in an extended argument with playtesters: that of having the same argument again and again over the course of years with customer after customer. I do ask myself: How strongly do I really feel about this? Is this a fight I am prepared to keep having? Sometimes the answer is yes. But here, where I had mixed feelings myself about how it was working, it frankly seemed like more trouble than it is worth. Since abandoning Confederate objective movement, I can’t say that I have ever regretted it. While it has made balancing the game harder in some ways, it has made it easier in others, and so from my personal perspective, it is a wash, and as I no longer have to hear playtester complaints (and fear the same complaints from customers for years to come), it is a net win overall.
As much as I would like to say that the problem of game balance is behind me, the fact is that it isn’t. While I would like to publish the game sooner rather than later, with a printed game, you only get one bite at the apple (software you can fix with downloads; wood and paper, not so much), and it needs to be right before the game can ship. My big fear right now is playtester exhaustion. The testers are only going to want to do this for so long, and then they are going to be ready to move on. Some turnover I can handle, but experienced playtesters are not easily replaced; new testers need some time before they really come up to speed and can contribute strongly to balance issues.
If you’ve been good enough to wade through the above fairly dry material on theoretical and technical issues of game design, you get a nice cookie as a reward: some photos of some of the playtest games. (Note: the positions are all from actual games, but were re-staged to make pretty pictures.) These pictures are from 4 different games, each one after the 4PM turn on 1 July and show a sampling of the variety of situations that can arise during play. As you recall, the motto for this design was “Death to All Plans!” and here you can see what that design motto means in terms of actual game play.