"Whoever made the remark about being accurate being like being pregnant - either you are or you aren't deserves a medal for Most Misleading Analogy and General Stupidity - even the most exact sciences now only talk in terms of progressively more accurate (and more complicated) approximations and this guy goes p******ing on as though there is some cosmic Answer in the Back of the Book and nobody gets any marks unless they guess exactly right."

Accuracy in game design is a subject that's always gotten its fair share of attention, and is getting more each day. As the hobby becomes more sophisticated, it demands more of its products, and since the hobby began with the claim (if not exactly the reality) that "now you too can refight..." it is hardly surprising that every major game designer spends more time than ever before worrying about whether he got everything right.

The question is, what does getting everything right entail, and how does one go about it? The second part of the question isn't too tough, once you know the first part, but the first part is a bit sticky.

Order of Battle

One element of game accuracy that receives more attention than most others is the game's order of battle, for two reasons. First, it is an element that is easily checked and one about which most people can come to some sort of agreement without a lot of subjective decision -making. Either the 15th Panzer Division was in Africa or it wasn't, and most of this information is a matter of record. A second reason for this emphasis is that it is a good general check on whether the designer knows what he's talking about. If the OB is hopelessly screwed up, or is exceeding superficial, one can assume that the general level of research that's gone into the game is likewise screwed up or exceedingly superficial.

Both of these are good reasons for giving the OB of a new game a good going- over, but this sort of thing can be carried too far. For one thing, the OB is not the only element of a game that contributes to its accuracy or lack thereof, (some would claim, myself among them, that it is far from being even the most important element) but it often is the sole element that receives any sort of serious scrutiny in a review, and reviews are just about the only type of articles dealing with the guts of game design these days.

This is especially disheartening when many reviewers don't even do a very good job of checking the OB's in games. Dishearting perhaps, but also very understandable. Although I doubt that anyone has pointed this out before, it is nonetheless true that reviewers and game designers alike are to a certain extent shackled by the same problems when confronting OB's - the lack of a clear and systematic methodology in the field of military history. OB's are included in military history to the extent that the historian is personally interested in that sort of thing, and not because there is any real pressure in the discipline to do so. Some historians actually frown on this sort of mundane "cannon counting" and so OB research tends to be catch as catch can for everyone interested in it.

It would seem then that no clear answer to the question of what constitutes accuracy is to be found in the existing body of literature (game reviews) on game design, which leaves it to me for the moment to put forward a tentative definition. I would begin by arguing that there are two distinct forms of accuracy: product accuracy and process accuracy. The former, if viewed in complete isolation from the latter, is typified by the "it all comes out in the wash" school of thought. If at the end of the game the Germans are at the gates of Moscow (or the French are on the field at Waterloo, etc.) then everything's all right. The latter, again in isolation, holds that the final outcome is irrelevant; what's important is the feel of each turn. On a microcosmic level, giving armored units big combat and movement factors is a sort of product-accuracy approach; it gets the job done, but doesn't tell you much about what's really going on. The innovation of the Mechanized Movement Phase in the Kursk/France '40 family of games, however, was one of the first real process-oriented rule approaches to what armor actually does.

Arguments as to which sort of accuracy is "better" are for the most part fruitless. While it is intuitively obvious that, if the process is accurately simulated the product must naturally follow, and while it is equally true that simulation of the product does not necessarily entail an authentic process, it is also true that it is impossible within the scope of a commercial hobby game to completely simulate the processes at work. Thus as a step toward a definition of accuracy, a game must include some elements of both process and product accuracy. Product accuracy must be designed in to take up the slack of those processes that are beyond the scope of the design.

The relative mix of product and process accuracy in a design is often what makes or breaks it as a game, but for me the answer to the question of how much weight to assign to either also provides the clearest general definition of what constitutes an accurate game. General Rule: all historical- based games are role-playing games. Each participant is cast in a specific historical role, and thus a game is most accurate if it confronts the player with decision-making problems similar to those which his historical counterpart faced.

Using this general rule provided both the designer and the player a good solid starting point for making decisions about accuracy. Nelson at Trafalgar has no business laying each gun of every ship in his fleet, nor is he concerned with abstract questions of ship design. Many first designs founder on the rocks of unnecessary detail mascarading as accuracy. If a player is confronted with a raft of decisions which have nothing to do with enhancing the accuracy of the role-situation he occupies, the additional variables add nothing but fatigue. By way of general conclusion, then, the job of the designer is to present an accurate "feel" of the situation treated by presenting the player with a detailed and accurate recreation of the decision-making process facing the player's historical counterpart, and abstracting those considerations above his authority or below his notice.

Logic

At this point, the logical thing to do would be to talk some about how one goes about doing that, but a different issue comes to mind that is potentially more important, and at least a little disturbing to me. In the process of thinking through this column, it occurred to myself and several of the staff that certain levels of games (notably true tactical games and grand strategic games) may by their very nature defy a treatment as accurate as is the norm for games in the operational vein.

In an operational game, it is fairly easy to roughly approximate the decisions facing a commander and limit his options to the historical ones. The information he has available can be limited and special cases kept to a minimum. In fact, historically, wargames have tended to concentrate on this operational level of treatment. Once a player is cast in the role of a head of state or an individual grunt, however, it becomes almost impossible to accurately present him with options short of including a referee. On an operational level, the commander at least has a fair approximation of what the process at work is. An infantry battalion may perform better or worse than expected, but the general characteristics of a battalion are known and can be approximated by considering things such as equipment, training, morale, etc.

The fundamental difference between this and both the true tactical and strategic level is that at those levels the uncertainties in decision-making have to do with, to use a game analogy, what the rules are. Examples might do a better job of explaining what I'm getting at.

In World War II, the possible reactions of the Vichy France government to several possible courses of action profoundly influenced the policy of both sides. In 1940 the Italians pressed the Germans for permission to occupy Tunisia and disarm the French troops there for the purposes of 1) obtaining a shorter sea route to Africa and 2) pressing into service the French equipment, most notably medium tanks and motor transport, desperately needed in Libya but not available from Italian production. The Germans refused since they felt this would prejudice their attempts to seduce Vichy France into the Axis coalition against Great Britain, a major German goal at that time. In fact fear of this eventually lead the British to bombard the French fleet in its North African ports. In retrospect, any damn fool knows there was precious little chance of Vichy France becoming an Axis ally.

Historically Impossible Alternative

However, is it more inaccurate to include such a historically impossible alternative or leave it out and thus not provide an historically very important motive?

Another example from WWII is that of Turkey. Britain intervened in Greece primarily to encourage Turkey to join the allies. The specific use of almost all combat troops available in the Middle East not absolutely necessary for garrison duties was due to the fact that a token action would have been recognized for what it was, and would have done more harm than good.

This raises a number of interesting questions, such as 1) What was the actual likelihood that Turkey would have entered the war on the Allied side at that time, 2) How much of a British success would be necessary to trigger such an entry, and 3) How do you write a rule that differentiates between a legitimate British commitment and tokenism?

Most of these questions defy answers, but the real problem is that answers to the questions do not solve the basic problem. The British did not know the rules regarding Turkish entry, any more than the Germans knew the rules regarding Vichy French entry, and the mere act of writing down the rules seriously distorts reality, regardless of how "accurate" the rules may be.

Very much the same sort of problem plagues tactical games, particularly those of a man-to-man sort. I would argue that one of the most important leadership variables at that level is the ability to judge character. Will Sam bug out or stick around? Should I use Garfield's squad as a base of fire, or Jones'?

You can perhaps simulate this in a truncated sort of way with random variables, but that merely imposes uncertainty without providing an accurate match of the correct type of skills of two people. Rules can be written that give one player or another an advantage in this area, but again this only dictates who will make the superior decision, not allow the players to give it a go.

In short, it would appear that only at the operational level can hobby games approach accuracy by presenting players with decision-making problems roughly similar to those encountered in the actual situation. If true, that is a very sobering thought, and on this subject in particular I would very much appreciate what the rest of you think. Provided we receive enough letters, I'll try to put them together into a forum and run it next issue.