Category Impact: The Dirty Little Secret About Wins
Don’t chase wins; they’re a crapshoot. We’ve all heard it and pretty much accept it as fact.
Wins aren’t nearly as much of a crapshoot as we’re led to believe. Let’s borrow a formula courtesy of Bill James to demonstrate.
James developed what he called the Pythagorean Expectation to estimate the number of wins a team should garner based on their runs scored and runs allowed. The formula generates the expected winning percentage and is as follows:
Winning Percentage = Runs scored ^2 / (Runs scored ^2 + Runs allowed ^2)
Instead of applying this to a team’s runs scored and allowed, let’s instead use a pitcher’s ERA and runs scored by his team to generate his expected win percentage. Then we can multiple the percentage by his starts to generate an estimate of expected wins. Finally, we’ll correlate expected versus actual wins to gauge the predictability of wins.
Before we break out the spreadsheets, the method is not perfect. All pitchers do not receive the same run support. In addition, a relievers contribute towards true ERA that should be utilized in the formula.
That said, this isn’t a thesis for SABR University or even a justification of a projection process. It’s merely a means to illustrate setting fantasy lineups in the hopes of garnering wins isn’t nearly the fool’s errand many lead us to believe.
To wit, here’s a table showing the correlation between expected wins and actual wins, broken down by ERA for starting pitcher tossing more than 50 frames last season.
|ERA range||# pitchers||correlation|
|2.50 - 3.00||14||0.95|
|3.01 - 3.50||42||0.91|
|3.51 - 4.00||42||0.94|
|4.00 - 4.50||28||0.95|
|4.51 - 5.00||23||0.83|
For those not familiar, a correlation of 1.00 means that one event leads perfectly to the other. A correlation of zero means the two events are completely random. The above results show a very high correlation between expected wins and actual wins. The next time someone claims wins are so ridiculously unpredictable, tell them Bill James’ Pythagorean Expectation begs to differ.
The reasons for the conventional misperception aren’t really germane to this particular bandwidth; perhaps I’ll pontificate on the topic down the line. What is relevant is you aren’t wasting your time if you consider wins potential when setting your fantasy lineups. This is especially relevant if you’re playing a daily game where the points for a win could mean a few extra Benjamins. Or Washingtons, if you’re playing the lower entry contests.
Since these are probably pretty obvious, we won’t spend too much time on stuff like target good pitchers at home facing lesser teams. Instead, we’ll suggest a tip well known to those dabbling in the daily fantasy arena and that is to look at the Vegas betting odds, especially if you’re searching for an under-the-radar choice for a win. This method isn’t perfect as the lines are set to equal the action on both teams, but they’re close enough so you’ll at least know who’s favored more often than not.
One of the caveats that comes with chasing wins is you’re usually choosing from options lower on the proverbial totem pole, which means the ratios that come along for the ride are often scary. We’ll broach this when we discuss the strikeout category since the same caveat will exist.
The important message when it comes to chasing wins isn’t so much avenues to pursue them; they should be obvious. The message is chasing wins is more than trying to get the pot at the end of the rainbow, finding a four-leaf clover or spotting a unicorn. The number of wins a pitcher totals in a season can be predicted pretty closely by the Bill James Pythagorean Expectation formula.
thanks for the pitchers' and catchers links' . appreciate that . however , this one is missing : "... here is how the 30 MLB teams rank in terms of stolen bases from 2011-2013."
I originally did it using IP and the correlations were even better (GS and IP basically measure the same thing) so I switched to ERA since that answers the question whether the quality of pitcher matters -- it matter a little but the correlation is still high.
Would be interesting to split by ranges by GS instead of ERA (e.g., 0-10 GS, 10-15, etc.) and see what the correlations were then.
Leave a Comment
Affectionately known as Lord Zola, Todd was the 2013 Fantasy Sports Writers Association recipient of the Fantasy Baseball Article of the Year, Web. He's been with Mastersball since its inception in 1997 and presently Todd writes for the ESPN Insider and Baseball HQ. Todd is a frequent guest on SiriusXM and is a regular on HQ Radio. He's a veteran of Tout Wars and LABR as well as a multi-time NFBC champion. Follow Todd on Twitter @ToddZola
- Teixeira in Yankees' starting lineup on Tuesday
- Padres reinstate Everth Cabrera from 15-day DL
- Werth not in the Nationals lineup on Tuesday
- Josh Gordon will dispute test results
- David Wilson forced out practice with neck burner
- Jay Bruce has been placed on the bereavement list
- Matheny: Believes Wacha will pitch again this year
- Romo (back) will participate in Tues practice
- Vance Worley Throws Complete-Game Shutout
- The Melk Man Delivers 2 HR in blowout
@asveenstra place-holder for a player not in database or to represent a minor-league/reserve draft pick to make it easier to track trades?
Having Josh Rutledge back on my @TheNFBC Main Event team is like having your dog run away only to show up at your doorstep two months later