Just updated the bout page to provide a lot more information for upcoming bouts and historical bouts.

The played bout version is much more helpful now by locking the resulting ratings to that point in time, as well as retaining all the prediction metrics from before the scores came in.

Also introduced here is the Expected/Actual DoS chart which is the underlying view of where the ranking system predicts a matchup, and where the resulting scores fall.  The farther apart these two ticks land, the more the ranking system will adjust the two teams' position.  If the ticks line up exactly, everything was as expected and no one moves.  See the algorithm docs for more information about DoS.

What's great about all this is it makes it very easy to see what the ranking system expected for an outcome, the actual result, and how that difference informed how the ranking system responded.

Remember you can get at this info by clicking any bout from the front page ticker, or by looking for the Bout Icon icon on various pages throughout the site.

The image below shows how the page looked before and after the Oly/Rocky game from this weekend.

Oly vs Rocky (Before and After)


Comments

We don't deserve such wonderfulness. Thank you, anyway. I want you all to know you've earn your way to heaven, for the awesomeness of your hard work.

For the 06/26/11 Arizona/Madison game there was a predicted ratio of 1:3 and post bout it is said to be 1:1... I have come to understand that FTS leaves one of the numbers when the final ratio is determined (this helps with comparison). In this case the wrong number was obviously chosen. If the 3 had been chosen then the ratio would have much more accurately been represented as 5:3

Hey thanks for the heads up. We have seen this in one other instance, and have a fix coming. We still want to frame it from the team that lost rating points to show how many more game points they needed to maintain their position. But since the expected score ratio was already capped out at 1:3, there's nowhere left to go for a whole number < 1.