Seminar CCXXII: counterfactuals and computer games

There were more seminar series at Birmingham than I could easily keep track of, and less well advertised than would have made it easy, but I was delighted nonetheless to see the name of old acquaintance and general nice guy Robert Houghton, now of Winchester, on a poster at some point in late 2014 and made a point of making it to the paper on 11th December 2014 even though I had no real idea what the seminar, the History and Cultures Workshop, was. It transpired to be one of the many postgraduate-run events and I think I surprised them by being a staff member (just about) there at all, but perhaps they were no more surprised at me than I was at Rob’s title, which was “Modelling the Middle Ages in grand strategy computer games”.

Game start screen for Charlemagne in Crusader Kings II

Not my screen, I don’t play, but this seems to be game-start condition for Charlemagne and you know, it could be a lot less accurate…

It transpired that Rob had, while teaching at St Andrews, also managed to find some extra work advising the makers of the grand strategy game Crusader Kings II—so if you play it and have been favourably impressed by its apparent accuracy that may be his fault—and the coincidence of his two rôles had made him start to think about how such games might be used as teaching tools, and whom that would reach, and he had started doing some very informal research among his teaching groups. After all, if it can work then games are an interactive and exploratory tool that are quite unparalleled by anything we can offer by more conventional means; I remember an admissions interview at Oxford in which the candidate was courageous enough to say that there was probably no better way of getting a sense of what it was like to walk around medieval Florence than playing Assassins Creed, and although I’m pretty sure it should be easier to kill yourself falling off the Duomo than it seems to be, I got and get their point.

Of course we are not complete strangers to walk-around visualisations in historical teaching, but it's obviously not as interactive as a game environment and you can hardly climb anything

Rob’s findings, anyway, with a very small cohort as he freely admitted, were more or less that most people don’t get their medievalism this way but for those that do it may actually influence their understanding very deeply. In that case it may matter what they are playing, because what’s available gives a rather Eurocentric, ultra-violent, male-dominated picture (although this is not unique to computer games, of course). However, there are limits on what can be done and still have a playable game. Particularly with grand strategy games the player has a level of abstraction, information and control that no medieval ruler ever did, but the gradation from there to single-handed sword swinger is very shallow. Also, the computer can’t create the environment without solid parameters. Where you don’t know something, you can’t just leave it blank; that pixel, that bit, must be present. There were other issues usefully pointed out too, both by Rob and the others present, but I got particularly engaged with the issue of counterfactuals.

The pocket battleship Admiral Graf Spee burning off Montevideo, 17th December 1939

The pocket battleship Admiral Graf Spee burning off Montevideo, 17th December 1939

Again I should preface this with a tale of an Oxbridge admissions interview but this time it’s my own. You see, at school, for the sins I mostly had yet to commit, I ran a paper-based wargame of my own invention for a year or so. I more or less stole the combat system from Fighting Fantasy and made it work for World War II and imagined-World-War-III air and sea combat. Because my path was probably already clear, I got interested in trying to recreate certain battles as tests of the game system, and especially the above, the Battle of the River Plate. I got it to the point where I could, with all kinds of reservation, make River Plate come out about right two-thirds of the time. Now, I must have mentioned this in my application form as most of what I remember about my Cambridge interviews is the late lamented Clive Trebilcock pushing me for five minutes or so on whether I thought such things could be used as tests for counterfactuals; if I could make River Plate come out right two-thirds of the time, for example, did that mean that this gave some basis for saying that the odds were genuinely against the Germans despite appearances? And I gave no ground: no, I said, because firstly I know the desired outcome and so can make player decisions that are likely to create it; also, because I can’t test for how important chance ought to be. I have set the parameters of the system—damage delivered by weapons, ability to minimise and inflict it, resilience of ships and crews—to give what seem to be like the ‘right’ outcomes, but the more accurate I get with that the less of a rôle chance has, and the more it slides from being a test of alternatives to just a recreation with dice. On the other hand, the more room for actual play there is, the less testably accurate it will get. I probably didn’t put it as eloquently even as that while 18 and nervous as hell, and there are ways I can now see to refine it—retests varying single factors, for example—but I still think it wouldn’t work.

Screenshot of a Crusader Kings II game

Not quite Rob’s example, but a situation in which something alternative has quite clearly and thoroughly happened whose story would be quite fun to know

More to the point, though, it wouldn’t be much fun if you could make it into a useful predictor system. And here Rob was on the same page. He had a marvellous Crusader Kings II screenshot in which someone playing as the ‘Duke of Alba’ had managed to conquer or otherwise gain control of basically all of north-eastern Europe and the Baltic. Now, I can just about think of ways something like that could be set up in fiction—royal Scottish exile during a succession crisis who has a monastic conversion and joins the Teutonic Knights, rises to be Grand Master and then gets recalled to the throne and, like Ramiro II of Aragón marries, for the sake of the dynasty, an heiress to Brandenburg whose brother dies soon after…—but it really is fiction, this could never have happened (even if ‘Duke of Alba’ were a real title). But should the game exclude it? Many reasons why not come to mind, not least trying to program for all such eventualities, but most obviously that one of the ways people treat computer games is to try and bring about heroically unlikely outcomes, winning through with the least likely playable character or from the weakest starting position and so on, and that this is one of the things that makes such games fun, because you can win against great odds. In Assassins Creed, I don’t believe it helps you particularly but you can climb the Duomo.

Screenshot of an attempt to climb the Duomo in Florence in Assassin's Creed

Go on…

Now, obviously sometimes in history people did win through against what seem to have been great odds, but as far as we are concerned as historians there was actually only one possible outcome of all the actual variables in play, and part of what we’re doing is trying to account for what seems to have been unlikely by identifying the significant variable. But even at age 18 I could see some reasons why a game will probably never be a tool for testing such outcomes; too much must be preset by assumptions about the outcomes. In some ways it might be better for teaching them, because then the presets are actually part of what we want to teach, but if the Duke of Alba can’t wind up ruling the entire southern Baltic coast or the Aztecs somehow reverse-conquer the Spanish Empire, then I’m still not sure I wanna play…

My only halfway reputable cite for using counterfactuals in medieval history is Jes Wienberg, “Europeanisation around the Baltic Sea: a counterfactual perspective” in Jörn Staecker (ed.), The Reception of Medieval Europe in the Baltic Sea Region: papers of the XIIth Visby Symposium held at Gotland University, Visby, Acta Visbyensia XII (Visby 2009), pp. 421-429, but that has a good go at defending the practice.

4 responses to “Seminar CCXXII: counterfactuals and computer games

  1. The older games of that particular studio (Paradox Interactive) had a wonderful capacity for being teaching tools. Unlike Crusader Kings II, they were not random. They helped me notice any simultaneous events around the globe – unlike textbooks, which are split into regional and themed chapters.
    Also, there were fans who were keen on extra detail. For a while, they kept modding these games to make a “hands-off playthrough” more historic.

    • Oh, well, there’s a perspective I didn’t have! I didn’t realise they’d done previous stuff and, now that I’ve looked, have never heard of any of them. But it is interesting to see this trend towards playable realism coming from the players, too. Thankyou for the comment!

  2. David Damerell

    I fear you don’t want to know how the Aztec Empire happened. It’s not a player getting super-clever, but a preplanned scenario where the Aztecs turn up with a massive invasion force and technological advantage. It’s about as good a counterfactual as “suppose Hitler has superpowers”, alas.

    Paradox games _can_ produce fascinating counterfactuals, mind.

    • Hullo David! That’s sort of disappointing, I admit, though I quite like the philosophical exercise in revising assumptions that it involves. And it’s probably fun to play, partly because of its madness no doubt, and therefore why not have it? They’re not trying to sell a simulator, after all…

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s