Ataxx & Clones scoring system motion

greenspun.com : LUSENET : MAME Action Replay : One Thread

BBH & I were debating of how we should score these games. And here's what we came up as a motion.

You start at ONE credit(and 100 seconds), at the easiest level.(level E, 800 rating).

If you win, you may feed another credit in and proceed at the harder level. If you lose, your game is over regardless of time left.

The score would equal the opponents rating * your score.

So if you play against the next-to-hardest opponent(2000 rating), and you scored 25 points. You would score 50,000 points for that game. Add the points up that way.

What do you think? GB9

-- Gameboy9 (goldengameboy@yahoo.com), May 28, 2000

Answers

This game may end up being a problem on MARP. It is a game I've taken much interest in. A simplified version was put into the PC game 'The 7th Guest' which got a lot of discussion at the time. I don't know if an ultimate strategy is yet known but several people have had a crack at writing software solutions/computer algorithms to solve this puzzle - I know I have. Some use brute force that 'look n moves ahead' where n is as large as memory/stack will endure while others use cunning variants based on reasonable heuristics. I anticipate that some people may have access to these programs and may begin running the solve program and feeding it the ataxx moves and basically facing off the AI's against each other. With todays cpu's, Ataxx's AI probably doesn't have a chance. A pity because Ataxx is an awesome game. I expect that fair play will prevail on a reputable site like MARP but we need to be aware of the possibilities. Over the next few weeks I'll be testing my AI program out on Ataxx, I'll let you know how I fair and how easy it is to 'face off' Ataxx against a computer AI.

-- Tim Morrow (tjmorrow@bigpond.com), May 29, 2000.

I'm a bit confused as to your scoring scheme. Just to clarify, do you propose that we have to play all opponents from easiest to hardest? i.e. Colony (E) rated 800, then Droolman (D) rated 1200, then Mush Man (C) rated 1700, then Gorgon (B) rated 2000, then Cephalo Man (A) rated 2200.

By score do you mean the number of blobs you have converted. Is your score summed over each of the 5 matches? i.e.

Score = sum over all matches i of score_match_i * Oppononent_Rating_i

If so then I don't mind. If on the other hand you want to score based only on the highest rated player played then I don't like the proposal as it stands. You might play well through 4 of the opponents and then get wiped on Cephalo Man with a score of 0*2200 = 0.

I would also like to see the remaining time be considered for points as well although I don't know how much weight should be given.

One anomaly that the above doesn't address is the early win. You might wipe the enemy by converting all his blobs before the board is half filled and not get much of a score. Extra credit should be awarded for being able to do this I think as it is particularly difficult to do. I think the score difference between you and the opponent is a better measure of performance.

-- Tim Morrow (tjmorrow@bigpond.com), May 29, 2000.


yeah i like the number of blobs you have times the opponents rating, i think time should only be added into the score when we make the game an all tg recordings game (so we can make sure no one uses pauses.) plus in ataxxj there is no timer that tallys the whole game, only one for a single move.

-- Chad (churritz@cts.com), June 10, 2000.

Wait a minute...

Here is what I want to know...

Why in the heck are we discussing an alternative scoring system, when the game itself has it's own scoring system. Just becuase someone doesn't like it can allow us to go ahead and replace it with whatever we like, if you beat a harder opponent, it rewards your points, if you lose, you go down, depending on how you faired. If you fair well against a tough opponent, then you can stay even or even higher.

-- Chris Parsley (cparsley1@hotmail.com), June 11, 2000.


The answer Chris is that we're avoiding a massive tie that we'd get with one credit. Even multiple credits can force a massive tie with this difficulty.

-- Gameboy9 (goldengameboy@yahoo.com), June 11, 2000.


gb, that wouldn't happen. You are allowed to choose against any of the 5 competitors, and the only way to effect a tie score would be to score the same EXACT number of pieces on the opponent... The game takes into effect the rank of the opponent you play, and how big (or little) you win (lose) against that player. I'll break down the formula the game uses later. Also it helps players who think quickly, because you can get two matches in under the 100 count on 1 credit. (Although we might want to think of switching this game to TGMAME only as soon as it's feasible.)

-- Chris Parsley (cparsley1@hotmail.com), June 11, 2000.

This is untrue, i just played Droolman winning with 31 and losing to mushman with 11 reds, and got 1241. then i won with droolman with 33 and lost to mushman 21 reds and got 1241. I should have ranked a lot more losing with many more reds in my second game but still got the same score. the ataxx ranking sytem does NOT take into account the pieces you win or lose by. which is why we need to add this to our scoring system to eliminate ties.

the ataxxj also does a similar thing but gives a higher score for winning droolman. 1700.

-- Chad (churritz@cts.com), June 11, 2000.


Moderation questions? read the FAQ