Sunday 14 November 2010

Campaigns part 2

My previous field of work was the fascinating world of Risk Assessment. Software that monitored incidents, identified key risks and trends. It's been used as evidence in court cases and SOX Sarbanes-Oxley 404 (I kid you not) compliant. This work gave me an understanding of visualising trends and the mathematics of risk and game theory. (Moving on, William Boyd style)

One of the early games I worked on was a little ZX Spectrum number in the 80s, I did the all singing, all dancing Amstrad CPC and Memotech (unreleased) conversions of Next War. Next War was originally a very complex mathematical board game with hundreds of easy to loose cardboard counters, maps of the Fulda Gap East West Germany region, rule books and curved graphs that covered every kind of engagement you could think of. It's a rare game to track down but here's a picture I stole from a collectors site...

The Next War - extreme 80s board game
Cramming that into 48k was a challenge, indeed it couldn't be done on the 64k Amstrad without paging memory so it ended up as an early 'floppy disc' enhanced game or 128k model. Still surprisingly playable and I have a Blitz version half-written somewhere.

Over the months when I've sat down to sketch how Combat-Helo was going to work as a game, we knew what the conflict was going to be, the scope, how it would evolve and who the participants were. It had to be smart, never quite the same game twice. Also it had to retain the charm of the old 90's combat simulations, have a seemingly simplistic scope but at the heart would be a hidden educational agenda.

War is something that is hard to understand. Seemingly incomprehensible and I don't, (I hope I don't) make light or trivialise it in any way. What do scientists do when they don't understand something? Well, quite often they make a chart from data.

My hidden agenda with Combat-Helo circa 2002 was an accurate number crunching simulation of resources vs. the mission (originally for the Basra insurgency at that time). The mathematics of war is an interesting topic that's become better understood with the availability of information, modelling and work done by scientists like Sean Gourley (Oxford Physicist) from whom I borrowed some concepts for Combat-Helo Operation Ouroboros.

Using methods of data filtering Gourley's team identified wars/conflicts have distinct signatures from which you can make predictions. (See Telegraph article). Start by simply plotting size of attack against frequency.  This line is also found in areas of Risk Analysis. These lines often comply with something known as "Power Law". Here's a graph stolen from Wikipedia.


Many natural and man-made incidents follow this power-law. Earthquakes are quite common but only a few are devestating, power-cut disruption vs population, size of craters on the moon and even the player rankings of any Call-of-Duty game. Players wanting to rank high require early adoption, more so in a hit based market, over time you'll end up in the same place regardless of game. It's one of natures rules. It wouldn't surprise me if Activision sales of these games followed the same trend, and an incentive to hike the price. This is something  suggested by recent pre-order sales boasts at Activision-Blizzard.

Insurgency and terrorism follow the same curve. With a few deadly events causing catastrophic damage with many smaller acts carried through. Military planning in the real world can use this to predict the frequency of the next big attack, this is something you'll experience in Combat-Helo. By changing the slope (a polynomial expression) we can change the level of insurgency around the map, and the slope can be altered by many factors which you can influence while in the air or your skills at planning in the command tent. Not only can this be applied to COIN ops (pun partially intended) but also armoured and open warfare.

Equation that predicts likelihood of an attack

What you'll have is a reasonably realistic level of enemy activity regardless of scale.

No comments:

Post a Comment