A discursive look at Napoleonic & ECW wargaming, plus a load of old Hooptedoodle on this & that

Wednesday, 9 February 2011

Computers in Wargaming - 2 - From Basic Principles

You've almost certainly seen this poster before. It has a nostalgic charm for me, since a version of this used to hang on the wall in the software engineers' office at our local IBM office, around 1978. The point, I think, is that it is easy to get so absorbed in the tactical requirements of a solution that you lose sight of what it was you were trying to solve. When I used to work with a bunch of systems analysts, we had a house rule that you had to start every piece of work by writing, on half a sheet of A4, what your objective was, and then keep the paper safely in your drawer. After that, you were supposed to take out the document once a week and check that that was still what you were doing. If it wasn't, you had to call a project meeting straight away.

I've been using microcomputers (as we used to call them) in wargames for so long now that I thought it would be a useful sanity check to try to reconstruct how I got where I am, and why. As anyone who has read my stuff before will expect, this is likely to require a lengthy ramble through my personal history, but - be fair, chaps - who else's experience was I ever likely to learn from?

In the beginning, the first wargames I actually organised and played (rather than watching from a distance at my local wargames club) were very simple ACW-based games, with relatively few troops and (almost certainly) the rules from Featherstone's Battles with Model Soldiers. They were fantastic - hilarious fun. In the company of some equally daft friends, I staged some of the least realistic wargames ever seen. Units of badly-painted Airfix soldiers whizzed around the table at motor-cycle speed, armfuls of dice rolled about the place, and the evenings passed amid a lot of laughter and a great deal of genuine (if rather childish) excitement. Every time, no-one would be able to believe that it was 1am already. There would - even at this early stage - be many occasions when we would forget to do something - usually, but not invariably, something fairly obscure, but sometimes something which completely derailed the flow of the game. One fairly common error was that, in the general excitement, one side would start fighting back or moving before the opponent had finished his turn. The resulting confusion could normally be rescued if spotted early enough, but it was sufficient of a problem to result in the drawing-up of some wall-charts to keep the play sequence organised.

Then, of course, my primitive ACW armies were replaced, first by Ancients, later by Napoleonics, and further reading and new rule sets (especially the preachings of the WRG) resulted in much thickening of the rules, more detail, and bigger and more impressive armies. The spectacle, of course, was improved, but the snags and minor irritations increased roughly as the square of the number of "improvements" we added. We tried to fix things, we placed more and more small print in the Commandments on the wallcharts, and, though we were still committed and positive, the games were never quite the same amount of fun again. We tried using a commercial rule set (WRG) exactly as published - soup to nuts - in the belief that it would hang together nicely, that 20 million flies could not all be wrong in their choice of diet. It helped a bit, but there was still a disappointing number of elements in the games which were - to put it bluntly - clunky, and it was far too easy to make mistakes which escaped immediate detection, spoilt the game and (for the first time) generated some resentment. That is probably an identifiable landmark - much of my devotion to the hobby ever since has had a lot to do with searching for the magic tweaks which might restore the joy of my earliest efforts.

The first serious realisation that something needed fixing came when, one evening, I found I was doing a headcount of the same unit for the 7th time in an hour, that I had once again forgotten to give someone the bonus points for their elite status, that the 28 of a possible 51 melee factor adjustments which I had just identified as relevant for an attack by the chariots on the left flank were exactly the same as they had been in the two previous attacks, and so on. And then came the creeping sophistication of victory condition testing each turn, weather, fatigue levels, progressive morale adjustments (as opposed to random morale throws when necessary), ammunition supply, national characteristics (oh God) - it became obvious that we were going to have to keep written records, so we moved bravely into a new era, in the belief that the hassle of the bookkeeping effort would be justified by a smoother, happier game.

And it didn't work very well. Now I know that many people use rosters to good effect, and swear by them, so this is entirely my view here, but - for me and my friends - it didn't work at all. For a start, the overhead of pre-printing record sheets and filling them in made it more like being at work than playing a game. Next, try as we might, we could not prevent the sheets of paper from cluttering up the battlefield and completely destroying the spectacle (and the spectacle, in hindsight, was one of a decreasing number of things we had left to enjoy) - from that point on, I have always loathed the sight of paper on the battlefield. Most irritating of all, the quality of record-keeping was so bad that it completely defeated the objective of making the game work better; worse, there was even a moral hazard here - if you forgot (or otherwise didn't bother) to update your roster for some loss or other, your unit strengths would be artificially high. Thus the sloppiest record-updaters would actually gain an advantage. Without employing independent auditors (and where has your game gone now, Johnny?), it just didn't work.

At this point some of my collaborators gave up and returned to their former life of visiting the pub, playing darts, even - it is said - spending time with their wives. The proper Old School doctrine would have been to go back to the original ACW games, make a feature of their primitive nature, and enjoy them for what they were. Somehow, life isn't like that. There is a faint echo of apples in Eden somewhere, but there is no going back. You can reconstruct the game and the circumstances, but not the innocence.

I had worked a lot with computers, and the very first affordable home computers were appearing. I commissioned a friend who had a Commodore to program some wargames routines for me, to see how it would work, and it looked promising. We did the game sequence (so that the computer announced each stage of the turn, and told you how many turns had elapsed), and we automated a simplified melee routine (for Ancients). Interesting.

I bought a Sinclair Spectrum, mainly because it was cheap and mainstream, and I started work to write a program which ran the move sequence skeleton, and which was gradually populated with more and more functionality, so that, over a period of time, a greater amount of the game was automated. It was a valuable experience to see it develop like that - you could judge whether each change was an improvement. A lot of it wasn't, and it was surprising how this worked out - something which looked like a sure-fire enhancement would prove to be just a nuisance, while some obscure, minor tweak might accidentally turn out to be a big step forward. We must remember here that these machines were agonisingly slow - I had to develop a good working knowledge of machine code to get some of the routines to work fast enough to be useful. One thing that never seemed a good candidate for automation was the movement phase of the games. To this day, in my computer managed games, the program does not know or care where the units are on the table, though it may know which ones have not arrived yet, or are not yet visible.

Around 1984, my wargaming involvement went on indefinite hold as a result of the pressures of the dreaded Real Life. One of the last things I did before storing everything away was to print out listings of all my Spectrum wargame programs - one of the smarter things I have done over the years. About 15 years later (Real Life having given up on me), I came out of the closet, and recreated the software, this time to run on an IBM PC. Apart from a lot more of the same, that is really where I am up to now.

It's maybe useful at this point to recap on what computers are good at, on what might be relevant and useful to a miniatures game, and on a few criteria and odds-and-ends by which we might judge whether the automation is beneficial. This is off the top of my head, so if you have a better list, please substitute your own.

Computers are good at

Storing and processing information, and reporting on it when you ask. Here is your roster, my Lord - and you don't even need to remove casualties.

Repeating a procedure, faultlessly, forever.

Doing calculations - of almost unlimited complexity - accurately, without forgetting anything and without making mistakes, and (sometimes regrettably) without coming back to tell you if this is a really stupid thing to be doing.

Generating random numbers - in more ways than you would believe.

Doing thankless background tasks - such as regularly testing for something that hardly ever happens - reliably and without complaining, and only telling you when the outcome is significant

They can also usefully do things without telling you about it - examples might be keeping track of the weather, or of concealed units (Blinds) or the arrival of delayed troops, or building a variance into scheduled events to provide an element of Fog of War - in an extended form, they can also go some way to providing support for a solo gamer - not by supplying a fully-functioning opponent, but by randomising things that you thought you knew, or by choosing one of a range of strategies, for example.

They can free you from some of the constraints of a manual game - for example, if your manual game uses a 6-sided dice to decide on some result or other, one obvious way to proceed is to automate it as it stands. In the manual game, a 6-sided dice is simple and readily available. On the other hand, a computer is just as happy with a 27-sided dice as with a normal one, so, if you know the game is to be implemented on a computer, you do not have to restrict yourself to the kind of dice which are convenient in the physical world. [In a business context, one of the most common inefficiencies in computerisation in the 1980s was the over-faithful automation of a clerical process, complete with all the double-checking, paper communication loops and other constraints which were inherent in the original clerical version. Sorry - that was boring.]

They can provide you with a Black Box game, which is especially sexy for a wargame - let's have a look at this:

Wargames, especially as they become more complicated, are hard to learn - there is a lot of stuff about how the troops behave, weapon capabilities, move distances - what we might term the historic aspects of the game, and then there is a whole lot of detail about how many dice to throw in various circumstances, exactly how many morale points to deduct if outflanked, all that kind of thing - the mechanics of the game. I get the horrors when I see a new set of published wargame rules which extends to (real example) 104 pages. Even if we ignore all the irrelevant photos of 54mm soldiers which have been used to brighten it up, there is still far too much in there. The author might understand it all, or at least he will know where to find the tables, but for anyone else this is a huge problem - especially for someone who, like me, regards a wargame as a social exercise rather than a plan to conquer the universe.

I am a big fan of the approach which does not require a newbie to know all the details of the mechanics, or what happens in the dark corners. In an ideal world, a wargamer with some experience and a good working knowledge of how (e.g.) Napoleonic warfare worked should be capable of learning the extra bits he needs to know to play a game in a very short time. The implication, of course, is that someone else - an umpire or a black box - knows the rest of it and makes the game work. Howard Whitehouse's Science versus Pluck rules for Colonial warfare follow this model - there is a very full, detailed manual for the umpire, and the other players only know what they need to know. The umpire obviously has a lot to learn and a lot of responsibility, but the situation for the players is what I would regard as correct. If they work, black boxes are easier to field (and drink less) than umpires.

One characteristic of a black box is that the contents have to be very carefully documented somewhere, and kept up to date as the logic is maintained. The program code contains a lot of wargaming nuts and bolts and expertise, and detailed knowledge of what is in there can dissipate very quickly. There are few less interesting tasks than reverse engineering the code to see what the game does (been there, done it...).

What sort of benefits are we looking for from the use of automation with a wargame? Well, as in every aspect of wargaming, what you are looking for depends on what you happen to like, but there are a few givens which I think few people would argue with:

The computer should not be a distraction, and should not impede, or detract from, the game it is supporting. The important bit of a wargame, after all, is the soldiers-on-the-table bit - the computer should be a help, but is otherwise not interesting in its own right.

The game should be easier with the automation, not harder, nor more irritating.

The means of input should not require a full time operator - even if someone wants the job (and would you invite someone like that to your home?), the constant passing of spoken information backwards and forwards is going to be an irritant, and, almost certainly, a source of fatal misunderstandings.

The program should be as failsafe as possible - for example, confirmation yes/no questions should be inserted as a double check at critical points, to guard against disastrous mis-keying, and (very usefully) a succession of security copies of the entire battle should be automatically saved at the end of each turn, so that the game can be rewound a bit if something goes horribly wrong.

That's probably more than enough for now. Next time I'll say something a bit controversial about user interfaces, and the way commercial games are designed, and I'll say a bit about games (including my own) of which I have some experience. There may be a little technical stuff in there. too, but only a bit, and only in passing. If you are still with me, then thanks, and well done!

If you are generally hostile to the entire subject of computers, there is some relief for you here.


  1. In no particular order:

    Yes I have invited such people to my house. Whether the game was better or worse for having them behind a computer vs actually playing is a state secret.

    I have, in the absence of funding for microchips or miniature RF transmitters, that a barcode for command stands along with a portable reader would be useful. Multiple networked smart phones might also work. (calling out data to said human interface definitely has been sub-optimal in my limited experience, not only due to errors but it interfered with the table talk which was sometimes more engaging than the games.

    Black boxes are good if you either don't know and don't care about what they are simulating but are willing to learn what works for that game, OR if the designer happens to agree with your interpretation of history, tactics etc. If one only occasionally plays the game, it can be deuced difficult to divine whether or not a given outcome was an expected result of your tactics because the designer believed, well lets say that columns should normally burst through lines or that cavalry could usually break a square or was the result of a low probability randomn number selection (not to mention a data input error 5 turns ago).

    Lastly, to my surprise, (this is the least relative comment of all), after lengthy contemplation I realized that the thing I missed the most when playing computer games, was the ritual and feel of rolling 6 sided dice, especially when one can read the results off the dice without needing to check charts etc. Its that emotional high when your hard pressed regiment rolls 3 5's and a 6 for their defensive fire as the enemy grenadiers come in (or the opposite).

    It surprised me because previously I had thought it was all about, firstly whether or not the game felt historically plausible and secondly if it flowed well and had decision points to maintain interest. That there might be an aspect related to stone age men throwing knuckle bones for stakes hadn't seemed applicable to me. I've been pondering what that actually means for about me for a couple of years now.

    Anyway, looking forward to the next installment.


  2. About as close to computer wargaming I've come in contact with is the Talonsoft Napoleon series. It really does remind you of a wargame and you can fight real of computer A/I.

    Maybe I'm a bit lost, but what is it you're wanting to accomplish by using a computer for the calculations? Maybe I'm different, but I don't think a computer does a very good job of doing the really unique things that happened.

    Maybe some Confederates got a hold of some bad Mexican gun powder and their muskets are misfiring a lot? Maybe one commander is having a bad day with his mistress and can't get his mind around the battle before him and so forgets to issue certain orders. So many possibilities that I don't see a computer's input as being useful for much other than dragging the game out longer.

    If you don't mind me asking, how much time do you think the computer use actually involves?

  3. Hi Gary - I understand that the Talonsoft games are good - I'm specifically not covering this type of computer/video game in these posts, not least because I know very little about them!

    The chance events you describe, some of which might be set up by a scenario, will presumably manifest themselves as a reduction in a leader's ability, or some unit's firepower, or something you can quantify and apply to the game. The rules either cover this stuff or they don't, independently of the presence of a computer. The computer only works with the rules it is given - it is an easy matter for the program to decide whether each of the commanders is having a good day. You can arrange for it to tell you about it or not, as you prefer.

    All of this sort of stuff uses random numbers - whether you use a computer, or dice, or you draw coloured beans out of a bag, it really makes no difference. Though I own some examples of other people's software, the only game-management programs I use were written by me, and are maintained by me. How much time they take up is an interesting question.

    If the game is designed in what I regard as the most helpful way, all movement is done in the old fashioned way, and the involvement of the machine is restricted to the players identifying a unit, or units, for combat (or anything you like), answering a very small number of simple questions about the situation (most of which will be y/n - in cover? general present? etc), then - since the computer has the rules and formulae programmed in, and since it knows all about the state and capabilities of the units, the result is calculated and displayed, and the rosters are automatically updated. When the computer asks 'any more combats (y/n)?' and you answer 'n', the game moves on to the next phase. So interaction with the machine is simple, but does go on throughout the game, so a proportion of the time will be taken up by this, but in many areas it may take less time than working with tables and dice. Some aspects of my own games (concealment, weather, for example) would be difficult to do at all without the computer. How much time do we spend reading rule sheets and multiplying a headcount by two-thirds?

    Hard to say. About as long as a piece of string. Writing and maintaining the software takes quite a bit of time between battles, but I enjoy that.

    As I hope to describe in the next post, most of the game-management software you can purchase is designed to use a mouse, and present-day standard Windows type screen layouts, and it's far too screen-intensive. In these games, unless they are played by a club who can deploy a full time operator (which can be awkward) for the games, then you are absolutely correct - a lot of time is spent fighting with the input screens, and diverting attention away from the actual battlefield. If you have witnessed one of these games in action, then I agree with you entirely - they are badly designed, and they take up a lot of time and effort. I like to think my games aren't like that.

    Regards - thanks for input - appreciated