I'm interested in getting feedback on the logic of a computer refereed Matrix Game. The next post shows the logic the Dylan Alliata is working on. I'd love feedback in the form of holes you can punch in the system.

I'd also like to hear about any other attempts people have made to set up computer moderated RPGs.

Chris Engle
Hamster Press

COMPUTERIZED MATRIX GAMES

I know it may seem impossible that Matrix Games (which rely so much on a referee’s judgment) can be computer moderated but I have a person in Washington DC working on just that. If this works, we can set up self running web pages that play games for all comers.

What I want to do here it to describe the flow chart logic of the program and see if you can poke holes in it.

The game is based on a web accessible chronological data base. The rules of the game dictate how new information is added to that data base.

The web page opens with a brief write up of a scenario (an opening, a cast of characters, a map, and a story hook to make players want to find out what happens next.) There will be links to examples of play, the Hamster Press web page (where they can get a free copy of the rules) and links to the various web communities, in other words a typical web page.

HOW TO PLAY

Below the scenario write up is a list of the last ten arguments to happen in the game. Below that is a list of the next ten arguments that might happen. Players can vote an argument out if they think it is illogical by clicking a box by it and pressing a vote/veto button. At the bottom of both lists is a text box in which players write their argument. Arguments will only be allowed to be one paragraph so they can be read and understood in the above lists.

Arguments submitted got to a holding place to be “rolled on” so that no more than ten new arguments appear a turn – a vital point so that the readers do not become overwhelmed by information.

Bit by bit the story emerges – very much like in a face to face game.

HOW THE GAME IS RUN

When arguments are submitted they are given a number in a chronological data base. They are also put in a holding area to be adjudicated. At set intervals (say once a day) the computer starts “rolling” for the arguments. Each argument has a 50/50 chance of making it through a round. Once the number of arguments has been cut down to ten the rolling stops and the arguments are automatically posted to the Vote/Veto section. If ten or fewer arguments are turned in during a turn then the computer does only one round of rolling.

Once arguments are posted to the vote/veto area people viewing the site get to judge if the argument makes no sense. Since the first round of dice rolling is essentially brainless, it can not filter out jibberish arguments. As players read arguments they click on a box marking an argument out as illogical/stupid/impossible/etc. When a new set of arguments are ready to move into the box the computer looks at the voting to see which arguments are vetoed. Normally it only takes one vote/veto to erase an argument. The computer now does a little cybernetic trick where by the player’s feedback tells the program who to judge the votes/vetos. If all the arguments are ruled out this suggests partisan bias. The computer then ignores arguments that only received one vote against. It also raises the bar if all the arguments are vetoed again. In the future two or more votes are needed to veto. This increases each turn up to say ten veto’s. Another feedback loop is if 3/4ths of the arguments are vetoed. This raises the required vote to veto to two. If 3/4ths are vetoed again it ups the requirement to three and so on but as soon as players start vetoing 50% less arguments then it drops the required number to one again. This makes highly partisan play more and more difficult but resets when players start being more reasonable.

Successful arguments are moved into the “Most recent argument” area and subsequently to the archive area. A game can be set to last a certain length of time or can be set to shut down if people stop submitting arguments. When this happens players have clearly lost interest so the game resets new.

ARTIFICIAL INTELLIGENCE

Is this artificial intelligence – NOOOO! Players can beat this system and break the game at will. It will only work if most of the players are reasonable which I believe is valid assumption. The players are the referee. They are the ones who make up arguments and decide which arguments don’t make sense. The computer only functions as an information clearing house.

Chris Engle
Hamster Press
4-27-05

Now that is kind of cool. However, I have reservations due to the points you raised about it being easy for players to break the system. Only testing would see how often that would happen though I guess. Regardless, it's a very interesting test of game design.

-ALEX

Thanks!

Dylan has some ideas on how to block spam jibberish makers. I'm not totally up on how it is done. If a player wanted to type jibberish it would take time which makes you wonder why they would do it.

The interesting possibility that this brings to my mind is the potential for these kind of sites to mushroom proliferate. A lot more people could play than I could run through a PBEM game. Which could bring people to playing live.

Chris Engle
Hamster Press

To be honest, thats what interests me in the idea. It provides a means for marketing via word of mouth (sort of) that can't be bad. Also I'm a web developer/designer myself so its intriguing to see my interests converge :)

-ALEX