In the fall of 2000, I got into what I thought was an interesting discussion about the nebulous, never-to-be-solved discussion about what constitutes *accuracy* in a wargame. Because it was off topic on the CSW For the People page, I moved it to the “Coffee Table”, where it promptly died. However, I came across these notes while looking for something… anything… to add to this site. Enjoy!
October 18, 2000
… I am fortunate beyond words that I have been able to select career paths that led to improving myself as a wargamer. (Just think of how bad I'd be if these hadn't have happened.) As a Marine officer I found out that controlling troops is a LOT different than controlling cardboard counters. It took me two years to get where I could reliably move a platoon 5 or 6 klicks without making (or letting them make) any serious errors. When I consider Grant's '64 campaign, I am amazed that he so nearly succeeded time and again in delivering a "killing blow" to Lee. This and many other "real world" lessons made me a great skeptic of all wargames and even more so of all histories. I can categorically state that THERE ARE NO ACCURATE HISTORIES, but there may be some accurate wargames.
I can say that there may be accurate games, because in the second part of my life, I became a professional wargamer -- an Operations Analyst -- and to avoid false modesty, a very good one. I know what goes into a real world simulation and more importantly what must stay out. I know that these "games" are not "fun" to play, and I damn straight know that decisions that I help make will determine what kind of vehicle or aircraft my son may have to fight in. It is from that perspective that I unequivocally second Mark's statement that the more that is left out, the better the simulation.
Not to bore you with “Theory and Practice of Operations Analysis 101”, but suffice it to say that "real" wargames are different enough from our hobby games that thankfully I can still use our games for recreation. (And incidentally Mark, who is also a professional Ops Analyst, can get his recreational jollies from designing hobby games.) Because it is a hobby, I can enjoy The Longest Day (aka the Longest Wargame) as a gruesomely detailed simulation of the Normandy campaign and just as readily enjoy Breakout: Normandy with its area movement, Risk-style combat resolution system. More importantly I could not possibly tell you which was the more accurate simulation… because they're both accurate. In one the designer leads us to manage replacements, logistics, artillery support and the apportionment of airpower. In the other we can focus more on the mass and maneuver at the Corps level… different designers - different perspectives.
Then Joel Tamburo replied:
November 3, 2000
I have been reading the posts on this board with a lot of interest. A couple of weeks ago, you posted the following remarks, which I and my friends found very interesting:
1) There are no accurate histories
2) There may be some accurate games
In the post, in fact, you mentioned The Longest Day and Breakout:Normandy specifically. Could you elaborate on these statements some more? Also, is FtP one of those games that may be accurate, and if so why?
November 9, 2000
Joel, you asked me to address (justify?) three things I said. I should warn you there will be a bit of semantic bait and switch. But in the end I stand by what I said because the roots of my point of view lie in the realm of mathematical chaos.
1) There are no accurate histories. (I'm limiting all my comments to military history.)
It is impossible for any history to be accurate in the strictest sense, because they must all be incomplete. Additionally historians are forced by their professional processes to rely on written records. Occasionally one finds historians willing to accept first person oral accounts but some historians look down their noses at this. Ambrose's recent history of D-Day comes to mind [Stephen Ambrose, D-Day, June 6, 1944: The Climactic Battle of World War II]. Not being an historian it seems silly to me as most written records rely upon unsubstantiated first person accounts. The best example is a unit's message log or command post journal. Many historians consider this to be a primary source as it comes directly from the command center involved and is (theoretically) untampered with. The problem is, even at the battalion level (the lowest echelon where such journals are usually kept), the reports received are fragmented and incomplete at best. Occasionally the messages received are intentionally erroneous. They are almost always interpreted out of context when read after the fact.
First person example. In 1979 I was a company executive officer during an amphibious exercise on the island of Viegas, Puerto Rico. The second morning after the landing I was rudely awaken by "friendly" aircraft bombing me. Fortunately these were pretend bombs, so the bombing was merely a loud and spectacular air show. About an hour later, my Marines spotted some soldiers moving along the road beneath our little mountain. After making sure there were no friendlies out and about, we deduced that they were "bad guys" and engaged them with a fusillade of blank fire from machineguns and mortars. A few hours later, I had business in the battalion rear area and as a matter of routine visited the battalion combat operations center to glean whatever information I could before returning to the company. First stop was the Intelligence Officer's situation map to check out the enemy situation. He showed an enemy company sitting on the hill that my company occupied. To shorten, what over a beer is an hilarious story, suffice it to say he and I got into a rather heated exchange over the accuracy of his map. He proceeded to "prove" to me that the hill must be enemy because they bombed it the first thing in the morning. On that we could agree. Finally, in exasperation, he thumbed through his copies of incoming messages and showed me the SPOT report (that I personally had sent) of my company engaging the enemy on the road. He had the pieces to the puzzle but had them put together wrong. And we weren't even in combat where things get really confusing! This is not an isolated or unusual case; feel free to ask any soldier/Marine.
The higher up the chain of command one goes the less accurate the picture becomes. Messages are delayed, lost, reinterpreted (corrected?), misrouted or simply intentionally trashed. It's amazing to me that Generals can control a battle at all. (The reasons that they can do so are interesting but not directly relevant to this point.) To get an accurate picture of what really happened, the historian would have to get a first person account from everyone in the battle (on both sides) and then put all those pieces together. Impossible of course. First, it is the nature of battle that some of the more important points of view are lost when their owner is killed. Secondly, after the battle the memory of every participant starts to reinvent itself to try to rationalize what happened. Read S.L.A. Marshall's autobiography or his Battle at Best to get an historian's description of how individual's accounts of the battle change quite rapidly after the battle. Compare General Schwartzkopf's and General Frank's accounts of the end of the Gulf War. Both are honorable men, and I don't believe either is lying (i.e. being intentionally deceptive), but their accounts can not be reconciled. How is an historian to figure out what happened?
He does the best he can to pull together an hypothesis that explains all the known facts and recognizes that it will be changed if and when other reliable information becomes available. Shoot, just think of all the WWII histories that had to be "rewritten" once the Ultra secret became known. What other secrets will never be known?
But all of this doesn't begin to scratch the surface of the problem. Combat is incredibly chaotic. In addition to chaos in the semantic sense, there is chaos in the mathematical sense. The slightest thing can make an immense difference in ways that are completely unpredictable. Shakespeare's lost battle for want of a horse captures this concept. But what about the grenade fragment that just missed killing Adolph Hitler in WWI? Think about it. A randomly fragmenting shard of metal probably of less weight than a dime that flew in a random direction and with sufficient random velocity to wound him but not kill him. It is the stuff science fiction, not history, is made from. But history is just one long conglomeration of chance events and imperfect decisions. How can written reconstructions be accurate?
>2) There may be some accurate games. (I'm referring here to well researched, designed and developed games. FTP is an example of this as are both Normandy Campaign games I mentioned).
This is where the semantic legerdemain comes in. I'm "recalibrating" (O.K. - changing) what I mean by accuracy. While we can't know, for example, the particulars of how Chamberlain's 20th Maine managed to win their desperate battle on the 2nd day of Gettysburg, we can accurately know that it did. We can speculate that this unbelievably heroic stand changed the course of the battle and therefore perhaps the war, but we can't know that. Maybe the 20th Maine folds but the Texans (no wait… Texans wouldn't have screwed up!) the Alabamans pull a Sickles-Peach Orchard maneuver, advance too far and are chewed up by Meade's arriving troops who then counterattack and collapse the entire Confederate front. Instead of ordering Pickett's charge, Lee spends the third day of Gettysburg trying to back his army out of the jaws of a revitalized Army of the Potomac. As Mark Herman has pointed out about a bazillion times on the ConSimWorld FTP Board, we can't know what might have been. In a mathematical sense, anything might have been if we allow all probabilities. As with Adolph Hitler's fragment, what would the Civil War have been like if Traveler had, while Lee was meeting the broken and retreating Virginians of Pickett's division, accidentally stepped on a spent cannon ball and fallen on The General breaking his neck?
In a sense games allow all these possibilities. Anyone who has played for more than a year can reel off examples where the highly improbable occurred. Chains of low probability events that, had they been written into a Hollywood script would be rejected as too coincidental, are frequently retold with glee. Yet that is exactly what happens on the battlefield. So a game inserts randomization into its story, and while a particular playing may be unlikely to the point of “inaccuracy”, the aggregate effect of all playing may be accurate. Like the historian, the game designer pulls together a hypothesis about what was important in a battle/campaign/war and what was not. He relies on the Central Limit Theorem (consciously or not) to establish a "probable outcome" and then describes the variation around that outcome to complete a CRT or Random Events Table or whatnot. As with history, paradoxically the game TENDS to become more accurate the more general and "strategic" its focus. This is another aspect of mathematical chaos; just as individual fingerprints are all different, they all still look like fingerprints. Even chaotic things tend toward a prototypical appearance and/or behavior. So too war.
What is fundamentally cool about good operations analysis models as well as good entertainment wargames, is that the designer can get historical truths (generalities) to manifest themselves by creating like conditions instead of specifying a particular outcome for a given cause and effect relationship. In the context of my original comments, this is what (I think) Mark was trying to say about wargame design. What you leave out is more important than what you put in… e.g. getting the right effect (as opposed to a specific historical result) from the natural interaction of the agents rather specifying every little detail. When this happens, the Operations Analyst (aka professional wargamer) can "discover" new truths, and the recreational wargamer can acquire new insights into why historical actors did the things that they did.
To my way of thinking, this ability to continually reexamine what happened (particularly if the same campaign is played in several different wargames) gives the "student" a much deeper understanding of what happened and what might have happened in actual history. A historian can come to an accurate general conclusion (Lee lost at Gettysburg) and a reasonable hypotheses explaining why (Stuart was missing, the Army of the Potomac was fighting in the North, the Army of N. Virginia was fighting in the North, Lee gave up the advantage of the tactical defense, etc. etc.) By playing a (set of) games we can come to an equally accurate, deeper understanding of the why. One designer might stress the effects of cavalry recon and screening, but another might stress the power of the defense over the offense at that time. When one has run tens of experiments in different games on the same conflict, some common elements will surface… those are most likely to be accurate, just as the things that are common in independently developed histories are likely to be accurate… and given the connections between wargame design and histories, they're likely to be the same things.
> 3) In the post, in fact, you mentioned The Longest Day and Breakout: Normandy specifically. Could you elaborate on these statements some more? Also, is For the People (FTP) one of those games that may be accurate, and if so why?
I'm still not happy were I left my last answer. I may not have discussed the mathematical foundation enough (and it's REAL easy to discuss it too much!) Fundamentally, USERS tend to rely too heavily on the accuracy of a history and discount the accuracy of a good wargame too much. I stress USERS because, I suspect that the vast majority of historians understand the inherent limitations of their craft, but many readers will read one book (usually one of those atrocious high school texts) and think they know what happened. When held to the same accuracy standard, e.g. read and/or play widely and from both sides, then games and books probably reach the same point. A student in that depth (regardless of media) will inevitably get a good understanding of what was involved and why certain things happened, what was true chance and what was "chaotically inevitable" and how everything seems to have fit together. More concisely, at that point both histories and games are equally accurate… or maybe inaccurate.
On to discussing the Normandy campaign games as an example. When I was but an egg, I thought that the more detail there was in a game the more accurate it must be. I now believe the exact opposite. The Longest Day (TLD) is a masterpiece. It has a superb map; IMHO the only better maps belong to the ACW series. (Those are easily worth the price of the game by themselves!) It probably has a better battalion level OOB than the generals fighting the battle had. (Actually it almost certainly does, as that isn't that hard to do, as I mentioned in part 1.) I don't have the math to back up this opinion, but I believe the battle resolution process and CRT combine to be very accurate; at least it "feels" so based upon my readings. As stated previously, every game designer crafts his game to capture certain dynamics of the conflict and excludes others. TLD takes a very close look at the integration and use of air power, the flow of logistics and replacements on both sides, the importance of weather, and to a lesser extent the seaward echelon as typified by the German Coastal batteries and the Allied Naval Gunfire. I've only played TLD in teams, but I assure you that the "Eisenhower" player, who must handle the air and naval fire support and push the resupply and replacements around, has as many decisions and activities as the corps commander players who are busy counting up factors and deciding where to attack or defend.
Breakout: Normandy (BKN) on the other hand has a regimental/brigade sized OOB and a grossly undetailed area movement map board. The combat resolution system is almost childishly "Risk-like" in its simplicity. BKN's designers (Stahler and Greenwood) also focused on capturing the importance of logistics, weather and the interaction between the two of them. Part and parcel of the logistics focus is the built in concept of culminating points as units become "spent". The game also ended up with a focus on bridges that I frankly find a bit distracting. I suspect they weren't so important during early prototypes but were included as the design matured. Their primary effect seems to be to "fuzz up" the boundaries between areas and in that function they serve very well. As with many other things handled by the players in TLD, the effect of allied airpower is greatly aggregated in BKN. There are two puny counters for bombardment that the allied player can use to mass effects. But the main contribution of Allied air power is its significant effect on German resupply and movement, and that is out of the players' hands.
If we were to compare multiple playings of each game, which "lessons" are common and which are different? The most important common lesson is clearly the tyranny of logistics on both sides. The allies were under extreme pressure to get adequate supplies across the beach to maintain forces in sufficient numbers to keep from being thrown back into the sea. It is possible for the German's to actually do that in both games. (In one of my proudest wargame victories, playing TLD, Panzer Lehr et al. first prevented an US link up at Carenten and then rolled up the Omaha beach Mulberry! The Allies were ham-strung and had no chance to win at that point.) Once the allies are getting enough logistics to keep from losing they must still get enough to start to attack and win. In both games, it is here that the contribution of allied airpower is felt. In BKN the linkage is simplified. The better the weather the lower the German supply. That's true in TLD as well, but there one can only isolate specific portions of the battlefield and never perfectly; again when the weather is bad enough, the Germans work like crazy to get well. Both games capture the commander's role in allocating his logistical support. With supplies and replacements so limited on both sides, the issue of which units are reconstituted and which are left to get by becomes critical. Few histories give these logistical issues more than a few paragraphs of coverage, yet a wargamer will come to not only appreciate them, but to live them. All of the sudden, the broad sweeping strokes and grandiose plans of the history reader must be put away as the player deals with the same urgencies of the moment that his historical counterpart had to deal with.
Another accurate (sic) common point is the use of supporting artillery. In both TLD and BKN it is standard practice to mass artillery from units not participating in an attack/defense to support those who are. This was standard doctrine in all the armies in Normandy. The effect, however, is to focus the battle even more tightly as both sides commit an increasing level of effort to a single point, e.g. an area or just a bridge in BKN and a sector of 4-5 hexes in TLD. At the tactical level I think both games do a good job of capturing the effects of the terrain. While in TLD one can move across the flooded terrain, after a few playings you quit trying unless you get stalemated. In BKN some boundaries are just left as impassable. The mythic quality of the bocage is handled well in both games at the operational level. It certainly helps the defender, but it is not impossible to cope with. In truth the game that probably captures the real effects of bocage the best is ASL where one can experience first hand its effects at splintering a once coordinated attack or defense into a series of almost unrelated small unit actions.
Obviously the largest difference is the level of detail. In a macro game like TLD the player(s) make a lot of decisions at multiple levels. I've already alluded to the allocation of air and naval fire support and logistics, but there are a multitude of operational decisions as well. In what area should I attack and to achieve what objective? Where can I conduct an economy of force defense in order to reinforce a more critical sector? Then once these decisions are made, there are tactical decisions of who specifically moves and which hexes are attacked and from which directions. Many players who are good at other games perform poorly at TLD because they can not segregate their operational decisions from their tactical decisions; probably more to the point they don't find it enjoyable to do so and that means that game is not for them. (It was a FTP discussion along these lines that started this whole ball rolling.) The BKN design, on the other hand, is focused much more tightly on a single level of decisions, the player manages divisions and regiments, their associated supply and integrates a limited amount of fire support. Missions are assigned by areas and are appropriate for the size units involved, e.g. go capture Carenten.
One more point (of a probably limitless number) to wrap up. Consider the victory conditions. In BKN the designer has "told" both players what areas are important. There is a VP level associated with each one; Carenten is worth 2 VPs, Caen is worth 4 VPs etc. However these same locales do not have VP associated with them in TLD. Yet in both games the battles swirl in the same places. In one case because that's were the victory points are, but in the other because that's where the roads are. Both are historically accurate; it is almost inevitable that Caen will end up contested. Typically the German will try to hold Bayeaux to prevent a link up but if he chooses to make a real effort to do so he will be too weak elsewhere and will lose much more in the long run. Carenten will be the site of a large and bloody battle where a German victory spells doom for the Allies, but the reverse is not necessarily true.
The student of military history and military operations can learn from playing both games that despite the heroism and sacrifice on both sides, the issue depended primarily on the unsung soldiers who marshaled the supplies and drove the trucks. He learns that von Rundstedt did not blunder by failing to prevent an Allied link up in Bayeaux; in truth there was little choice. Perhaps he learns that Montgomery wasn't the boob that some American histories make him out to be. If he uses the game to explore other strategic options, he can "discover" that Bradley saved the landings the previous December by rejecting the initial plan and insisting on adding Utah beach and another Airborne Division to the American sector.
Which game is more accurate in my opinion? Probably BKN because it limits what it is trying to accomplish.
Which one teaches the best lessons? That would depend on the student.
Which is my personal preference? TLD, because I happen to like putzing with all those pieces and worrying each little decision. But NOT because I think that the detail makes it more "accurate" or "realistic" -- it doesn't. In fact, I think TLD is realistic in spite of its design, not because of it.
Where does FTP stand? For its domain (a strategic level simulation of the Civil War) it is far above TLD or BKN (in their domain) in both playability and fidelity/accuracy. FTP is chock full of insights, some of which none of us (including the designer) have even suspected.