Thanks for playing

[Cross-posted at Revise and Dissent.]

Niall Ferguson has an article out in the New York Magazine, on the use of computer wargames in learning about history and strategy. (Via ClioWeb). It's a frustrating piece. As a sometime wargamer myself, I do agree with him that they can have their uses. But I think he fundamentally, and strangely, misunderstands what those uses might be.

Ferguson first outlines why he sees a need for such games:

Today we live in a multipolar, multiplayer world. Some players are much better armed than others. In that sense, today's strategic problems are more like those of the World War II era. Sure, the U.S. can invade Iraq. But what will the French do? The Russians? The Chinese? What if invading Iraq ends up benefiting Iran? The question is, where to learn this kind of stuff?

He thinks wargames can help with understanding the big picture, and he may be right. In gaming a battle or a war, things like terrain, supply, position, maneuver, firepower, and so on move from being abstract factors found in books, to concrete parameters that need to be juggled in order to achieve victory. Ferguson (quite fairly) dismisses computer games like Civilization and Empire Earth, which include WWII scenarios, 'since what they provide is such a crude caricature of the historical process'. He then looks at (what he claims is) the previous best attempt, the board game Axis & Allies:

Up until now, the best my sons and I could do when it came to replaying World War II was in fact an old-fashioned board game, Axis & Allies. Similar in its mode of operation to the earlier strategy game Risk, Axis & Allies offers a reasonable approximation to the strategic position in 1942. But I stress approximation. The game vastly understates the economic power of the United States, for example. The best thing about Axis & Allies is that battles are decided by a combination of firepower and luck. Dice are thrown, but the odds are weighted in favor of the player with the most men and hardware. (Each time I play, I'm impressed by the calibration of these weightings.)

Now, I know Axis & Allies well. There's just no way that it is the best simulation of the Second World War there is, or even was. It's not much less of a caricature than Civilization, or Risk. The individual pieces represent -- well, who knows, but they do not vary between nations in either strength or number: Britain can, in theory, have as a big an army as the USSR; German aircraft carriers are just as good as Japanese ones. The world is divided up into irregular-shaped areas -- "Eastern Europe" is one, stretching from Istanbul to the gates of Leningrad. But it can hold no more units than can Guadalcanal; an infinite amount, actually (but only one AA unit). Production of new units is governed by "Industrial Production Certificates", which you get every turn depending on the territories you hold. But any you don't spend in on aircraft production in 1942, say, can be saved up until 1945 to buy tanks, as if industrial production was some fungible thing that can stashed away in a warehouse until it's needed and turned into whatever is required.

I don't want to bash Axis & Allies too much; it wasn't intended to be a hyper-accurate simulation, but just a fun game. (And it is -- it's hard to not to enjoy pushing those plastic tanks and battleships around the world!) And even so, it's not completely divorced from reality; maybe in its cartoonish abstraction it's still teaching its players useful things about how the war was fought on the largest scales. This is certainly one of Ferguson's points, but then he turns to a new computer game which he thinks blows Axis & Allies away, The Calm & the Storm. For one thing, 'it is based on a quite astonishing quantity of factual information about the war':

Like Axis & Allies or Civilization III, the graphic interface is a map. But the level of detail is quite unique. Not just national borders but provincial borders are visible. And all the world's countries are depicted; players can choose from up to eleven governments, including China's.

Here Ferguson shows his limited knowledge of wargames, for there's nothing particularly 'astonishing' or 'unique' about the level of detail described here, even when the discussion is limited to games covering the Second World War at the strategic level. For example, the map for The Calm & the Storm is much coarser than that for World in Flames, which covers much of the world with a hexagonal grid: each hex is 100km across on the European map, and 230km on the Pacific map. And while eleven playable countries is impressive compared with most board wargames, it's not when placed next to another computer simulation, Hearts of Iron II, which has dozens of playable countries, from Turkey to Tannu Tuva.

Ferguson is also impressed that diplomacy is a crucial part of The Calm & the Storm. Again, games like Hearts of Iron and World in Flames (via its prequel, Day of Decision) were doing all this years ago. But this leads to a much more serious problem with his article. (After all, there are worse sins than not being a hardcore wargamer!) He uses The Calm & the Storm to explore a counterfactual history of 1938:

I argue in my new history that confronting Hitler in 1938 would have paid handsome dividends. Even if it had come to war over Czechoslovakia, Germany would not have won. Germany's defenses were not yet ready for a two-front war. So how did my preemptive strategy stand up to a computer stress test? Not as well as I had hoped, I have to confess. The Calm & the Storm made it clear that lining up an anti-German coalition in 1938 might have been harder than I'd assumed. To my horror, the French turned down the alliance I proposed to them. It also turned out that, when I did go to war with Germany, my own position was pretty weak. The nadir was a successful German invasion of England, a scenario my book rules out as militarily too risky.

This is the part I have a hard time understanding. Why does Ferguson trust the results of a computer game over his own historical research and judgement? Wargames aren't magic; they aren't somehow automatically correct, as though all you had to do was program in the laws of war, feed in some data about numbers of tanks and production of ball bearings, and then crank the handle to see what comes out. (Although some might disagree.) They are no better than the historical knowledge used to create them, and possibly worse, in that they often involve assumptions and speculations to cover the gaps in what we know about the past -- a wargame designer must of necessity make judgements that cut swathes through decades of studied academic ambiguity. (So do historians, at times: but wargame designers are not obliged to show their reasoning or their sources.)

So, Ferguson's reverence for wargames (or at least, this particular one) undermines his argument that military historians should be using wargames as research tools:

'What if D-day had gone wrong?' is only one of scores of counterfactual questions historians have asked about the war. What if the Nazis had invaded Britain in 1940? What if Hitler had captured Moscow in 1941? What if the Japanese had won the Battle of Midway in 1942? These are questions that computer games ought, in theory, to be able to help answer. And yet no military historian, to my knowledge, has made use of them. This is doubly surprising. Not only is there a long and respectable tradition of war games within the military academy, but games also played a central role in Cold War strategy, advancing an entire branch of mathematics -- game theory -- in the process.

If historians are going to use wargames for research purposes, then they need to thoroughly understand the models the games are based on. From the article, it is not clear how deeply Ferguson has delved into how The Calm & the Storm actually works. For all we know, the algorithm for amphibious landings might be too generous to the invaders, or the code implementing it might be buggy. And there's no reason to believe that Muzzy Lane's game developers have any greater insight into France's willingness to go to war in 1938 than Ferguson, or anyone else. Do they have access to primary sources that the rest of us don't? It seems rather unlikely. So why he apparently privileges this game over other forms of historical inquiry is a mystery to me. I suspect he is blinded by the sheer density of detail in this particular game, and being unfamilar with the wargame genre, overestimates its novelty while underestimating the potential problems. Creating a wargame is largely an act of historiographical interpretation, interpolation and extrapolation, and any historian following Ferguson's lead would be well advised to understand this.

Investigations of a Dog also critiques Ferguson's article, from a different angle.

Update: another critique, from a fellow former WiFer, and current professor of literature, at AmericanHeritage.com.

CC BY-NC-ND 4.0 This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License. Permissions beyond the scope of this license may be available at http://airminded.org/copyright/.

20 thoughts on “Thanks for playing

  1. Alun

    It's an interesting problem. When you see the legions marching in Rome: Total War it's tempting to ask what you can do with them, but you're right about needing to understand the basis of the model to work with it. I don't know if the modelling is at the stage where you can start asking more than some blunt questions about some historical episodes, but it could be a way of organising complexity and the interaction of many different elements. I'd expect the model would break down, but often that's it's greatest use in showing where an assumption doesn't work. Producing a war game as an output rather than an input is an intriguing proposal.

  2. Why should anyone be surprised that Ferguson's wealth of military expertise is based on a board game slightly more sophisticated than Risk?

  3. Let's be clear that there's a difference between what Niall Ferguson actually thinks and what Niall Ferguson thinks will make an attractive and quick article... The point of this is to let everyone know that he's writing a book about the Second World War, and some of the 'radical' views he'll put forward in it (probably derived from a quick scan of Williamson Murray's _Change in the European Balance of Power_, it seems likely), not to suggest that wargames are necessarily a way to judge interpretations of the past.
    To critique the content may therefore be unjust, but being unjust to History's Jeremy Clarkson is fair and appropriate. It seems to me he makes a category error when he assumes the game is accurate because it's hard to rearm as Britain, and that therefore it is a worthy judge of whether the French would join an alliance for war in 1938. But pace of rearmament is comparatively an easy thing to model as opposed to the dynamics of personalities and assumptions in diplomacy.
    He's also just plain wrong to say that military historians don't use tools such as these. Isn't this _exactly_ what Stephen Biddle does?

  4. Kurt Niehaus

    There is a big difference between computer models and Games. Scientists of all flavors have been using computer models for decades-- arguably one of the reasons they invented the computer. The best models are those that seek to answer a specific question that can be simplified with equations. Those questions dealing primarily with diplomacy probably cannot, while those dealing with production and logistics will see much better results. If one looks at any good editors notes for a game, the trade off between realism and playability is always mentioned. One of my favorites is from Squad leader/Advanced SL. There had been a problem with players voluntarily setting fields ablaze and burning cities to smoke out a few squads of opponents. The ability was real, and effective -- if shortsighted. This ability was removed as it was not widely used during the Second World war. Examples of things often overlooked include Ammunition usage and time constraints. Ammunition usage in most games is very much simplified, but in reality it has been one of the most important features in modern warfare. As for time, it is remarkable how much or little some units can accomplish, as there is difficuty trading functions that take time, In other words, can a single person run 100 yds under fire, find a target, radio for support, rally men, and keep himself appraised of what is happening in his unit in 6 minutes? 10? If all he does is run, how far can he go? Is he tired when he gets there? HOW tired? Can he recover? Does he get a short temper if he has been under fire for 30 minutes and cannot get radio contact? I think simulations might be able to help historians with understanding many things, or serve a springboard for further investigation, but not "games."

  5. History's Jeremy Clarkson ...

    Ouch!

    Shame Ferguson has no real knowledge of wargame history - no mention of Avalon Hill's Third Reich, for instance, which makes Axis & Allies look like noughts-and-crosses, and which has a very sophisticated economic-diplomatic-military model. But as Dan says, this article is less meant as a serious inquiry about gaming as it is a plug for Ferguson's new money-printing machine, I mean book.

  6. By the way, a quick bit of self-promotion; I wrote the introductory blurbs for the Great Powers in Hearts of Iron. I could never get into the game personally though (I much preferred its predecessor, Europa Universalis). Now I'm trying and failing to wean myself off Civ IV ...

  7. Post author

    Firstly, I should just say that I've never really gone in for Ferguson-bashing. It's hard for a blogging history-type-person such as myself to criticise another historian for writing extensively for the public. He does seem say outrageous or silly things, maybe just to be thought-provoking, but then so did A. J. P. Taylor (and I nearly chucked The Origins of the Second World War across the room few times!) I've only read a couple of his books but at the time I liked them -- where they touched on areas I knew something about they seemed quite good. I seriously doubt he's getting his understanding of warfare from Risk or even Axis and Allies. I think he's too smart for that!

    But as Dan suggests, if Ferguson has said something silly, for whatever reason, then he's fair game for criticism. Here, I actually don't think he's plugging his new book, because (a) he only mentions it two or three times, and more importantly (b) he doesn't mention its title anywhere! If he's trying to sell anything, it's The Calm & the Storm, which he mentioned by name many times, praises fulsomely, talks to the CEO of the developing company, etc. Perhaps I'm naive, but I prefer to take Ferguson at his word and assume that he's naive when it comes to wargames.

    Some good points from my fellow wargamers :) Kurt's right to point out some of the deficiencies in wargames as simulations. Often these come down to making a game that is playable, and enjoyable. Or sometimes it's because of inherent limitations in the components or software. Often it's not clear whose viewpoint of a war or battle you are supposed to have. I think most wargames are best suited as teaching tools, than as research ones. They at least get you thinking about various factors, such as the geography, forces available, etc, even if you don't take the course of gameplay as gospel truth. As Kurt suggests, you would probably need models, rather than games. Maybe with a monte carlo approach, running the war through thousands of times to map out the phase space ... a WOPR for historians! The Dupuy Institute does something like this, how sophisticated or sensible I don't know. I also didn't know about Biddle's work, that Dan mentions, is that his book Military Power? As Alun notes, a wargame could be a research output instead of an input ... but I suspect the beancounters would have a hard time figuring out how to much weight to assign it in your productivity and impact measures!

    Finally, Alan -- we are not worthy :) I have to admit that I also preferred EU to HOI2, partly because by the time I got the latter I was into my PhD and don't have the time or the energy for such a heavy game!

  8. '...because (a) he only mentions it two or three times...' ROFL.

    I should say that I have nothing against him personally (or Clarkson, come to that), and I actually think Niall Ferguson fulfils a very useful public function, much as Taylor did, which is to be provocative and make both the general public and the academy think about history. His books on banking are really good, and the stimulus of his book on the First World War had a very positive effect on my field as a whole.

    But there is a problem that the cat you try to put among the pigeons can end up becoming a pet - readers take a highly opinionated view as fact. I think that to write with such certainty about history is rather to conceal its complexity and difficulties from the public - which is not really about communicating a sense of our profession to a wider audience.

    Stephen Biddle's work - both Military Power and his articles on war in Afghanistan - are well worth seeking out. Biddle uses a range of techniques to seek out the key components of military power - including archival research, but most interestingly for this debate a Defence Department computer simulation capable of handling enormous numbers of variables. I don't think that his historical research is always _absolutely_ watertight (bit too much reliance on OHs in his chapter on Operation Michael) but the willingness to change significant factors and see what impact they might have had is very interesting, and well backed up with plenty of clearly presented statistics. I haven't done it justice here - have a read.

  9. Kurt Niehaus

    I don't know how effective the Monte Carlo sims would be. I know that West Point has been using company level command simlations for years. I actually had a chance to use it.
    The first question for a computer model is what are you trying to model. (some pretty funny Dilberts have resulted by not wanting to ask this question). For a Historian, I could understand trying to find a better explaination for a cause. Why did the French crumble so in '40? When the tanks are compared, the French have a good chance, based on strength of numbers. How do you account for morale? How do you account for more German Radios? I think the value would be much more in MAKING the sim than in actually running it. Unless I'm misunderstanding the Monte Carlo, the only effect a Monte Carlo would have is to determine how far off "expected" the actual result was. That is, given 100 Barbarossas, how often do the Russians win? Is there anything to "tweak that helps them -- say not having the red packets? Another thing that could be improved with simulation is understanding the "Fog of War" That's the biggest thing I got out of the WP sim. When I started losing tanks on my left flank and couldn't tell where the fire was coming from, and my 3 sub-level commanders were all screaming for the 2 helicopters and 8 arty guns, you realize how crazy the decision making gets.

  10. Post author

    Sure, monte carlo isn't a panacea. It would not explain why particular decisions were made, for example (e.g. for political or ideological reasons). And wargames obviously correlate far more closely to traditional military history than they do to war and society studies. But I think that playing around with monte carlo simulations might lead you to ask different questions. E.g., if it turned out that the Soviets won Barbarossa 1 time out of 10, then what were the key events or decisions that would need to be different? And if you can tweak things to improve their chances, then you could ask why they didn't do those things. The thing is that playing a game through once or twice, like Ferguson did (on the basis of his article, anyway) is hardly going to give you a statistically valid basis for conclusions on what is or isn't a likely outcome, even assuming the game is basically accurate. And my gut feeling is that if it isn't, you're not likely to get much more out of designing it, either, than you would by conventional historical methods.

    Of course, ideally there would be different historians developing different models, comparing the results to see which is best, and improving the next generation models, and so on ... It would be a lot of work and given all the uncertainties, maybe they would get just as much out of playing Axis & Allies :)

  11. Jason

    Brett, this reminds me that I have your Axis and Allies game on a shelf here... just by the way!

    As a comment, for what little it's worth from a non-historian, it strikes me that in many cases even the more sophisticated historical simulations like World in Flames and the like need to have special rules for those highly singular events. For example, there are special rules in WiF about unit placement by the US prior to a Pearl Harbor-style attack.

    Similarly it is hard to imagine a generic rule about atom bombs that would guarantee a Japanese surrender as historically occurred, but also capture the reasons why Berlin wasn't nuked as well.

    My first undergrad year included a few history subjects, including one that was exclusively about WW2. The course began with a map exercise in the tutorials - marking different places on a map that ranged from easy (France, Italy) to medium (Lithuania, Estonia, Stalingrad), to fairly hard (Rabaul, New Britain springs to mind). I was the only one to get them all right (except for a Baltic mismatch), just because we'd spent the preceding summer planning and executing world domination in WiF. The lecturer asked how I knew all those places, and when I answered he commented that he had had a few students previously who showed an unusual level of knowledge due to war games.

    BTW I am not actually claiming to have an unusual level of knowledge any more, but it certainly helped at the time.

  12. Similarly it is hard to imagine a generic rule about atom bombs that would guarantee a Japanese surrender as historically occurred, but also capture the reasons why Berlin wasn't nuked as well.

    Although I think your general point is sound, this actually isn't a very good example, because IMHO the only reason Berlin escaped the A-Bomb was that it wasn't ready in time. Certainly no moral or racial considerations prevented the city's almost total destruction by conventional weapons.

  13. Jason

    Although I think your general point is sound, this actually isn't a very good example, because IMHO the only reason Berlin escaped the A-Bomb was that it wasn't ready in time. Certainly no moral or racial considerations prevented the city's almost total destruction by conventional weapons.

    I think that's a fair point.

  14. Post author

    Your specific example aside, Jason, I think you're exactly right. Many decisions in war-fighting are not made for purely military reasons -- some good examples are Hitler's various operational interventions. The various no-retreat orders, for example. Or even better, his decision to declare war on the United States -- something nobody playing Germany in a game like WiF would ever do while at war with the Commonwealth and deep inside Russia, unless they'd gone stark raving mad. WiF gets around this by making a Japanese declaration of war on the US a declaration of war by the entire Axis. I'm not a US historian but it seems to me that it wouldn't have been a trivial task for FDR to get Congress to declare war on Germany if Hitler hadn't obligingly got in first. So those rules are there to entice players into making this same decision for which there is no strategic, in-game, reason to otherwise.

    About wargames being a great way to learn lots of factual stuff like geography (the AmericanHeritage.com writer mentions that he came away from WiF knowing the names of every capital ship in the war), that's true, but I also think that probably works for certain personality types, shall we say :) I suspect that some of our fellow players from that game didn't come away with quite so much knowledge! And of course, if you don't have enough knowledge of the real WWII to contextualise the information from the game, much of it would be worthless or even misleading. (E.g., the US entry stuff.)

  15. Nabakov

    Just a quick reminder that Len Deighton's 'Spy Story' makes a few interesting observations about frisky and feral wargamers' ability to puncture cherished tactical assumptions.

    "He's landing Be-10 Mallows on the ice to drop sonar through the icecap. He can't do that!"
    "He just did. We always wondered why the Sovs kept a squadron so far north."

  16. Nabakov

    "I'm not a US historian but it seems to me that it wouldn't have been a trivial task for FDR to get Congress to declare war on Germany if Hitler hadn't obligingly got in first."

    Perhaps not trival but FDR would have made it happen anyway. Let's not forget he was the smartest and craftiest politician aiming to be world leader since Palmerston.

  17. Post author

    I agree, it would have happened -- eventually. But I think the delay would have complicated Anglo-American strategic planning and might have made it politically difficult for FDR to pursue a Germany First strategy. Hitler should have been aiming to make things difficult for FDR, not make them easy. I guess the other thing, though, is that it gave the U-boats a big opportunity to savage American coastal shipping. I don't know whether that was a consideration for Hitler or not.

  18. Pingback:

Leave a Reply

Your email address will not be published. Required fields are marked *