This is our old blog. It hasn't been active since 2011. Please see the link above for our current blog or click the logo above to see all of the great data and content on this site.

## Tutorial Videos: WPA Graphs

Posted by Neil Paine on March 24, 2010

Today, I show you how to use the Win Probability graphs in box scores:

### 3 Responses to “Tutorial Videos: WPA Graphs”

1. Dave Says:

Quick comment:
What if you could take the same theory and apply it to the season for a team?
I mean nobody wins or loses a World Series based on the first play of the season but it would be fun to see how a team's pressure builds in terms of seasonal philosophy.

Maybe only for a divisional title...?

2. eorns Says:

These are great! Is there any way to find games that had the biggest or most swings, like the most times one or both teams blew 95% probabilities of winning? Sort of like the longest length of the line on the graph...

3. Andy Says:

Dave---coolstandings.com does that sort of thing. It's not that they use outcomes of individual plays, but they chart each team's chance of winning their division, winning the wild card, etc, from the beginning to the end of the season. I don't know very much about their probability engine but it does seem pretty accurate to me. For example, come August and September, games between the first- and second-place teams in a given division produce larger swings than games against non-division opponents or games earlier in the year.

That last statement may seem deceptive--it's not that games later in the year count more in the standings, but rather once earlier games have already been played and fewer games remain, each game carries more weight in determining possible changes in the standings as the season ends.