Hey Tango...who runs fangraphs (and how do I contact them?)

[SABR Matt]

Hey Tom...

You are familiar with my little thumbnail game by game pythag tool (PythagenMatt)...but I'm not entirely satisfied with the system, though it is a statistically significant upgrade over seasonal pythag.

If two teams go into the bottom of the 19th tied at 0-0, and the home team finally scores the decisive single run, that game should read as essentially 50/50...but PythagenMatt gives the home team one full win and the road team ZERO wins for that game (if you don't score, you have zero percent chance of winning by Pythagoras).

If one team scores scads of runs in one horrible late inning bullpen meltdown, but the game was close until the collapse, it should not read as utter domination by the winning side, but all PythagenMatt sees is the end of the game.

What I would like is the average win probability for each game. fangraphs.com does not currently track this, but IMHO it would not only improve game by game pythag substantially, it could be a revolutionary tool for rating strength of schedule in the modern era.

Do you know who I would contact to suggest that fangraphs look for a way to track and display that data?

[Tango Tiger]

Average win probability? Interesting. So, a team that starts at say .500 and ends at 1.000, the average will probably be something like .650 or something (as they oscillate between .300 and .700 in the early innings, and then march toward the 1.000 in the late innings).

That's kinda neat. If you remember, we once talked about how best to measure the "exciting" games, and I figured the best way was simply by Leverage Index. However, your proposal here would seem to also have this benefit as a byproduct. The more the game is in doubt, the closer your average win expectancy would be to .500.

In any case, the guy you want is David Appleman, and he has a message board here (registration required):

http://www.fangraphs.com/forums/

And a blog here (no registration required):

http://www.fangraphs.com/blogs/

If you still need his email, contact me offlist. But, you'll likely get a better response by using the above two.

[SABR Matt]

Not sure average win probability would do what I was thinking it would for improving game by game Pythag...but it would be an interesting thing to look at nonetheless (measuring the intensity of the games, giving you an idea of why teams are beating Pythag...etc0

I'll drop by his blog and ask if he'll add that info.