The formulae are almost identical, except for the correction of some programming mistakes the Zone made. There is a very minor difference involving changes in ratings when a player is new.  Also, Warp Gammon does not round ratings to the nearest integer, as the Zone does.

Warp Gammon has a rating system that is updated as soon as a result is recorded. Hence as soon as you have reported a result in the Quantum your ratings are immediately updated. If you mistakenly record a result then when it is amended, the ratings are stepped back so that the new ratings will be correct.

Ratings for in-room tourney results are updated at the finish of the tournament.

The main factors affecting your rating are these:

Match length  - the longer a match is, the more it is worth.

Rating difference - when a higher-rated player plays a lower-rated player, s/he risks more and gains less. The opposite is true for the lower-rated player.  The greater the ratings difference, the greater the risk/reward ratio for the lower-rated player.

Experience - when you are new to the rating system, your rating will be more volatile with each match showing a greater effect than it will when your experience is greater.

Some examples

3-point match, 150-point difference.  If the favorite wins, s/he will gain 2.95 points and the underdog will lose the same; if the underdog wins s/he will gain 3.98 points and the favorite will drop by that amount

7-point match, 250-point difference, 7.21 vs. 3.37

11-point match, 100-point difference, 7.88 vs. 5.38

25-point match, 300-point difference, 16.98 vs. 3.02

While ratings may fluctuate up and down, points are a measure of cumulative achievement and cannot be lost. Points are awarded each time you win a match, the number of points depending upon the event. In addition you can win bonus points winning a place in the event or bracket. Major events will have the opportunity to win gold points. 

A new player is normally given a shadow rating of 1500. However, when a player has an established record this may be manually set higher. As a player's experience increases, his shadow rating will be blended with his actual rating.

The weight given to the actual rating will be .5 ^ (E/600) for E<=1500, and, after experience equals 1500, the actual rating alone will be used. So after 600 experience points, a player's shadow rating will be a 50/50 weight of his actual and shadow, and after 1200 it will be 75% actual and 25% shadow.

We do not hope to identify every player in every tournament who should have a non-1500 shadow rating. Our judgmental shadow rating standards are as follows:

Top 10 in the world:   1950 (Nack Ballard, Neil Kazaross, Fallaffel)
Giant 32 not in the top 10:   1900 (Paul Weaver, Steve Sax)
Established tournament strong performers:   1850 (Phil Simborg, Gregg Cattanach, Ray Fogelund)
Top eschelon Zone players:   1800 (RedTop, wuffers, Doublefour)

There are two elements to the rating formula.  One is the likelihood of each player winning, the second is the number of rating points at stake.  For example, consider two players playing a 9-point match, where A is rated 200 points higher than B. Using the FIBS formula, there are 12 points at stake, and if A wins he will gain 4 and B will lose 4; if B wins he will gain 8 and lose 8.

If the number of points at stake in a given match changes, this will not change the long-term rating of a player. It will only make it more or less volatile. In the above example, if A and B are rated correctly, and the formula is changed so that there are 6 points, or 30 points, at stake, A will still, in the long run, remain rated 200 points above B.

It is important, however, that each match properly reflect the likelihood of each player winning.  Say that A and B play a 25-point match. Under the FIBS formula, A is 80% likely to win, with 20 points at stake. If, however, the actual formula used gives A 6 points when he wins and drops him 14 points when he loses, rather than the correct gain 4/lose 16 ratio, A will go up in rating and B will go down, if they play a large number of 25pt matches.

If the odds of A winning a 7pt match are 60%, the odds of his winning the series can be determined by adding the following numbers:

60% * 60% * 60%
60% * 60% * 40%
60% * 40% * 60%
40% * 60% * 60%

These correspond to the possible outcomes of the series:

W W W
W W L
W L W
L W W

These give a series-winning expectation to A of about 64.8%

Below we show the odds of the better player winning a 7pt match according to the FIBS formula, the odds of winning a best 2-of-3 series based on the formula above, and the odds of winning a single 16pt match:

50	46.2%	44.3%	44.3%
100	42.4%	38.8%	38.7%
150	38.8%	33.4%	33.4%
200	35.2%	28.5%	28.5%
250	31.8%	23.9%	24.0%

Thus, the odds of the better player winning a best 2-of-3 of 7pt matches is almost exactly the same as the odds of winning a single 16pt match. The same is true of a best 2-of-3 series of 5pt matches compared to a single 11pt match, although the correlation is a little less precise.

Accordingly, Warp Gammon will rate best 2-of-3 series of 7-point matches as a single 16-point match, and best 2-of-3 5-points as a single 11-point match. These will result in ratings for these series conforming to the aRatingsopriate risk/reward ratio that the actual series have.