AGEOD Forums

I think that it would be a good thing to have a method of measuring a player's strength in AACW. In my considered opinion, it would benefit players, the forum community, and AGEOD. If ever there was a good time to start such a thing, this tournament is it.

I'm thinking that there should be two types of ratings: Provisional, and Standard. The difference between them is that a Standard game factors in each player's rating, while a Provisional game does not. I suggest that a player's rating remain Provisional until they have played sets with three different opponents. An untested Player's provisional rating begins at 1000. I propose that playing strength ratings be established for AACW based upon results of two game sets, where each player plays each side of a scenario once. Both games of a set must be played, before ratings are adjusted.

There are three different types of scenarios in AACW: Battle, Two Theater, and The Grand Campaign. I suggest that provisional ratings be developed by playing Battle scenarios which have "sudden death".

I propose that these games be hosted by neutral moderators who are active in the forum as demonstrated by rank, and who have an understanding of the game demonstrated through reputation. (I suggest Lieutenants and above who "know their trade".) Exceptions could be granted for long standing forum members with low post counts. The basic idea here is that the person isn't going to quit participating in the forum, anytime soon. (Of course, exceptions to this rule could be granted for the current tournament, since most of the current hosts are involved in games of their own... and probably not going anywhere, anytime soon.)

I propose time controls be specified for each set. Maybe provisional games have a move rate of six moves per week, and four moves per week for standard games. This is just a guideline; players might want to agree on a faster or slower pace. If the neutral moderator agrees, then its all good.

I've developed an equation for ratings, based upon the primary elements of a victory. Not in any particular order, those elements are: Victory Points, Sudden Death, National Morale, Casualties, POWs, the duration of the game, and Player strength rating.

Here's my proposed rating equation for provisional games:
Points = National Morale bonus * Sudden Death bonus * VP ratio * Designed Duration

The winner adds and the loser subtracts Points to/from his or her current rating.

Here are the terms of the equation:
National Morale bonus: Winner's NM / Loser's NM
Sudden Death bonus: Designed Duration / Turn of Victory
Combat Losses Modifier: (Player's Combat Losses - Enemy POWs) / 1000
Adjusted Victory Points (Adj VP): Player's VP - Combat Losses Modifier
VP ratio: (Winner Adj VP / Loser Adj VP)


For Standard rating contests:
Winner's Rating Modifier: Opponent's Rating / Player's Rating
Loser's Rating Modifier: Player's Rating / Opponent's Rating

The Winner's Rating change:
Current rating + (Winner's Rating Modifier * Points)

The Loser's Rating change:
Current rating - (Loser's Rating Modifier * Points)

At the end of the two game set, a player's results are averaged.

The concept of player strength ratings for AACW has been inspired by Chess ratings, but that's where the similarity ends. Unlike Chess ratings, this ratings system tends to reflect the margin of victory, and rewards a potentially significant bonus for sudden death victories. "Quiet" games have fewer points on the line. As a rule of thumb, a game which is essentially a draw is going to have "Points" that are close to being equal to the number of turns.

On the other hand, violent conflicts that end in sudden death will cause an impressive increase in the victor's rating! (And of course, a hair-raising plunge in the rating of the vanquished. )

One of the cool things about this is that with enough data from enough rated players, we'll be able to give Athena a rating, for various combinations of difficulty level and options! Looking further, this can be used to help newcomers figure out their playing strength. We'll be able to make a recommendation of AI settings for a particular scenario, and ask them to play both sides. With the results, we can calculate their provisional rating... and so, they'll have an idea of who they want to play in their first PBEM.

I'm including a spreadsheet so that you can plug in the data from some of your previous games and see how you would "rate".

Dixicrat wrote:The concept of player strength ratings for AACW has been inspired by Chess ratings.

But..... we will earn as money as chess players????????????????

Manstein wrote:Turn 18 (final)
pkpowers (USA) / Manstein (CSA)
Morale USA / CSA
100 / 100
VP USA / CSA
159 / 167

Now, we will begin the second game, pkpowers as CSA and Manstein as USA

If you'll provide Combat Losses and POWs for each side, I'll be able to calculate a preliminary Provisional Rating for y'all.

Dixicrat wrote:If you'll provide Combat Losses and POWs for each side, I'll be able to calculate a preliminary Provisional Rating for y'all.

USA casualties: 210
CSA casualties: 1680
USA PoW: ??
CSA PoW: none

I didnt keep that game so I cant really say.