THE BOOK--Playing The Percentages In Baseball

One thing to be remember when considering Vegas’s “accuracy” as it figures out who is good and who ins’t: Vegas isn’t trying to predict winners. It’s trying to predict how people will bet. Ideally, the same amount of money will be bet on each team, and the books take their 10% from the losers. They have no interest in who wins when they set the line.

Dana: we’ve had several threads on this issue.  And the conclusion I came up with is that it’s something like 90% setting the true line, and 10% figuring out how much action on each side, that that’s the optimal way to set the line.

It’s a fascinating topic, and you can check out the archives for it.  I think the most recent one we had was 2-4 months ago.

Dana, go to post 11:

http://www.insidethebook.com/ee/index.php/site/comments/does_vegas_know_how_to_set_the_total_runs_over_under_line/#11

Thanks, Tom. I’m new here (just a couple of weeks), so I wasn’t aware of the prior discussions. I’ll check them out.

BTW, in case you’re wondering how people end up here, I found you from reading the Bill James web site.

How much good/bad luck did the top/bottom pitchers on the list have this year according to DIPS/FIP, etc. and compared to relative run support (adjusted for team offensive strength)? 

I’d expect the top guys on that list were relatively lucky, and the bottom guys were relatively unlucky compared to league average.

Dana: no problem.  Bill James once said something like:

If it seems like you are walking into the middle of a conversation, it’s because you are.

***

So, I know it’s tough to keep up, especially if you are new around here!  There’s over 6000 threads according to my logs, and at least 1% of those threads are important.  Problem is finding those needles.

(I prefer seeing things as 1% full, and not 99% empty.)

King: excellent idea!

So, there’s another project for you aspiring saberists (where are you guys hiding?): take the top 20 and bottom 20 in Xei’s list (*), and tell us what their BABIP is, tell us what their run support is, etc.  How did they perform with men on base v bases empty.

That is, find out if there’s something about them that would bias the results.  If there’s nothing in there, then that means that Vegas was a bit slow to recognize these pitchers.  (Not just Vegas, but likely all of us.)

(*) You guys have great handles!  King, Xei… great names.

#7. Would be an interesting exercise.  Btw, looks like the link in the article needs fixing. 
vr, Xei

I save daily data for SP’s, so I was actually going to look through their xFIP-ERA at points throughout the season and see if these extreme cases did benefit from good luck.

What made me think of this was seeing Greinke up at #7.  He was one of the pitchers this year that was always bound to regress back from his 5+ ERA and I don’t recall him ever having a line worse than -130 until August or something, even with an xFIP in the 2’s all year.

Tango, I have pitcher’s season to date data from June through August. Obviously sample size comes into play with only half a season+ worth of data, but perhaps combing that with there preseason projected FIP (I think I pulled Zips projections?), do you think that would give a better indication of who was lucky than only looking at end of season data?

Kevin, here are the lines (converted to win percentage, easily reversed engineer to ML Odds) for all 162 games for the Royals that I posted over at Royals Review yesterday.  The starting pitchers for both teams are listed and it is sorted by favorable odds for the Royals.  You can check on your Greinke hunch there. 
vr, Xei

http://sbn.to/stJEzP

Sorry Kevin, old habits hard to break.  Greinke is of course on the Brewers now.  Here is the Brewers chart (Fanpost at Brewcrewball)

http://sbn.to/uBGfwj

Tom,
Thanks for the link ot the other Vegas betting post. That’s exactly what I was getting at, and looking for.

One thing I have wondered about is how Vegas sets its line in the first place. How important are the sarting pitchers on each team? Do they account for half the odds or more or less? Let’s say team A is 25% better at hitting than team B but the starting pitcher for team B is 25% better than team A’s starter? Would Vegas call this even or give the edge to one team? 

Maybe it would be better to put this in terms of runs. Team A scores 4 runs a game while team B scores 3.5 runs per game. The starter for team A allows 4 runs per game and the starter for team B allows 3.5 runs per game? Does Vegas call this even or does one team get favored?

I am thinking how Vegas sees the question of “baseball is ____% pitching.” If it see it as 50% (or maybe pitching/fielding), then I would expect the games I describe above to be even odds (maybe I need to set the numbers up differently because if we ran them through pythagoras we might not get .500). But if they see pitching as more important, it seems like they would sets odds to favor the better pitching team.

cyril/13

The starter only pitches about 6 innings. So pitching is 2/3 starter, 1/3 bullpen.

But if the two teams are even in terms of expected run differential with these starters (along with their everyday bullpen, hitters and defense), then they should be set at even odds.

In other words, a run given up is worth exactly the same amount as a run scored, so baseball is 50% defense and 50% offense, but the starting pitcher is less than 2/3 of the defense.

And Vegas odds do skew a little bit toward public perception (and the public might slightly over- or undervalue starters, I don’t know), but they tend to be pretty close to the true odds of a team winning most of the time.

Thanks. So it looks like the odds makers see baseball as 50% pitching/fielding.

The guys that set the initial line don’t know any more than sabremetricians.  There is no mystical figure in Vegas that knows the odds better than the guys on this blog (including many of the posters).  Once the line is posted, the betting starts and smart guys often reshape the line by betting heavily on what they perceive as incorrect prices.  The sportsbooks then move the line.  This continues until game time.  At the end, the opening line that the bookmakers put up may or may not be close to the closing line.  It is the closing line that matters, not the opening line.  And thus, it is the market that dictates the line, not the Wizard of Oz.

One thing I have wondered about is how Vegas sets its line in the first place. How important are the sarting pitchers on each team?

I’ve wondered about that myself.  The best way to study it would be to look at games where the one of the initial starting pitchers is scratched after the initial line is set, but early on gameday, and see how the odds change.

LV (and internet where most of the money is bet) lines in short:  Pinnacle Sports hires smart people and they put up an early line.  Pinnacle allows betting with smaller than normal limits (normally $500).  Smart bettors bet.  Pinnacle knows who the smart bettors are, and they adjust their line as reqd.  Note that Pinnacle doesn’t care if more total money is bet one way or the other.  They note which way the majority of their pros bet.  “Everyone” copies their lines.  If you doubt this, watch Pinnacle’s lines move, and then watch other sites to see if they follow.

The short answer is that people (the net flow of money offshore and in Vegas) tend to value starting pitchers too much in isolation.  That is, if a guy projects to be 3.30 ERA pitcher ("ace"), and the rest of the staff is 3.40, he will often be overvalued.  If a guy projects 4.20 ("grinder") and the rest of the staff has been 4.70, he will tend to be undervalued.  People have a tough time separating pitching from the rest of the team performance; if you can do it correctly, you will beat baseball—even betting closing odds.

#19. I’m not really sure what you are trying to say here.  Certain types of starting pitchers are over/under valued in the betting lines?