Welcome back to another installment of Sabermetric Saturday, our weekly look at the best of sabermetric research, writing, and commentary from the past week.
Optimal launch angle off bat, by park | THE BOOK--Playing The Percentages In Baseball
Really cool graphic from Greg at HitTracker, showing optimal launch angle to hit a home run...and then showing, at least for Fenway Park, that significant park factors exist. It's better to hit high flies than line drives at Fenway if you want to hit a home run. Like a lot of things this weak...yes, not surprising, but it's nice to see the data.
How Much Does a Strong Division Hurt a Team? | THE BOOK--Playing The Percentages In Baseball
MGL tries his hand at strength of schedule assessments. He finds that the range of schedules varies by ~+-2 wins. Most teams, of course, are somewhere in between. The Reds' strength of schedule hurt them by ~3 runs or 0.3 wins. I'm surprised they fall into the "hurt by schedule" group, even if only slightly, because of the weakness of the NL Central.
Great data and presentation by Jeff Zimmerman breaking down trips and days on the DL by position. Pitchers account for 50% of all DL trips, days, and money. This is actually less than I expected, given that pitchers are fragile and account for ~50% of the active roster.
The Baseball Analysts: Comparing Division Projections
Dave Allen gives us one of the best presentations of team win projections that I've seen. Graphics tell the story, and Dave's as good as they come with a figure. In the NL Central, it's pretty clear that most systems see the Reds in a second tier below the Cardinals with the Cubs and Brewers. Houston and Pittsburgh make up the bottom-dwellers. I can handle that.
Pre-Introducing Batted Ball FIP | THE BOOK--Playing The Percentages In Baseball
In a very lengthy post, Tango demonstrates that there is such thing as a HR/FB skill based on actual variation in HR/FB data among pitchers compared to expected variation. This runs contrary to previous thought, so it's a surprising finding. It also lays the groundwork for an alternative way to evaluate whether a given item is a skill beyond the typical y-t-y correlation or intraclass correlation studies we've seen previously. I'm still getting my head wrapped around it, but it seems really solid. Anyway, the immediate application of this is that it does make sense to explicitly include a HR/FB term (perhaps regressed--you need ~600 air balls to regress only 50% toward the mean), in a DIPSy stat. Tango is in the process of doing just this (you can find an early version here)
Visual Baseball: Introducing the Paintomatic | The Hardball Times
Interesting new visualization technique to try to report pitch selections by pitchers. They show frequency of a pitch, as well as the average run value of the pitch. The latter, it should be noted, is a function of how often the pitch is thrown (see the next article by Sky A, or this one by MGL). This is why Tim Wakefield's blazing 72 mph fastball is shown as a "filthy" pitch--it's unexpected when he throws it, and so it's a good pitch under those conditions. So, there's a sense in which I think those data aren't particularly meaningful (as much as we'd like them to be!). I still like the pitch frequency visualization, though.
The Baseball Analysts: Hitter Scouting Reports
Sky Andrecheck has a great look at the pitch selection and pitch run value data at fangraphs, from the perspective of hitters. He finds a) that variation among hitters in pitch run values is almost swamped by random variation, especially for anything offspeed, and b) that variation in pitch selection given to hitters is very much non-random. Jay Bruce received only 54% fastballs last year (MLB avg 60%), compared to Willy Taveras's 65% fastball selection...'cause, you know, why not give Willy a fastball?
Speaking of pitch selection, I put together a simple game theory model to help explain the relationship between pitch frequency and pitch effectiveness/value. That information can be used to help formulate an overall pitch selection strategy, i.e. how often should you throw a fastball vs slider. It's undoubtedly oversimplified and can't yet arrive at quantitative predictions, but it's a nice demonstration of the logic. How's that for an unabashed plug?
The Baseball Analysts: Whose Stuff Plays Up?
Really interesting work by Jeremy investigating how pitchers change in velocity and pitch selection when moving between starting and relief. We know that if you use the same pitcher as a starter, his stats will be worse across the board than they would if he was used as a reliever: more K/PA, fewer HR, and even a lower BABIP. It's a huge effect, close to roughly a full run of ERA on average when moving between starting and relief. This is why you see a lot of failed starters achieve success in the bullpen. What we don't know is why this is, and that's what Jeremy's article helps us learn. He finds pitchers add ~0.7 mph in relief, and throw fewer offspeed pitches. What's really interesting, though, is the large spread among the pitchers--not everyone seems to react in the same way.
The effect of pitch sequence on batting eye and selectivity | The Hardball Times
Craig uses his batting eye statistic--a measure of how well a batter can determine whether a pitch is in the strike zone--to look at pitch sequences. His findings will shock you: batters are better at judging a pitch to be a ball or strike when it's a repeat of the same pitch! Ok, maybe that's not shocking, but it's nice to see the data. Also, you should throw a change-up after a curve, a breaking pitch after a fastball, and a fastball after a fastball (don't know what to make of that one). Of course, if you start to follow these patterns, the hitter will catch on, and negate your advantage. Darn you game theory!
Dave compares CHONE and FANS projections at FanGraphs. For several months now we've recognized that the FAN projections seem to be overly optimistic compared to other projection systems. Essentially, with fan projections, we have a Lake Wobegon effect, where (almost) everyone is above average. Looking at the data Dave provides, it looks like the biggest offenders are actually the above-average players (greater than 2 WAR)--apparently fans don't know about regression to the mean! Below-average players seem to be forecasted fairly close to the mark, most of the time.
Bill tries his hand at a catcher pitch framing metric: how many wins do catchers who are good at inducing called strikes earn for their teams vs. catchers who are not so good? Bill finds...an absolutely enormous, repeatable effect: plus or minus 6-7 wins. WINS, PEOPLE. Discussion in the comments largely mirrors my reaction: a) that is almost impossible to believe, and b) ok, maybe it happened, but we can't attribute all of that to the catcher. But even if we're plus/minus 10 runs (1 win), this is a very impressive skill that I've always dismissed as not that important.
Colin takes a short and sweet look at small ball events in MLB...and finds no indication that small ball is back on the rise. I think, in general, folks think small ball is back simply because scoring is down compared to a few years back.
Sky Andrecheck presents his Fan Strength Index, which is basically how much better teams did in attendance compared to what you'd expect based on team wins, playoff appearances, stadiums, etc. It does not include market size. Anyway, it shows the Reds as being a fairly middling franchise, never quite on the same level as other teams in the NL Central. I did very similar work a few years back, though I did include market size estimates: http://www.basement-dwellers.com/2006/03/quantifying-fan-interest-pt-3.html Because of that, however, I wasn't able to do the historical look that Sky does, which is really neat.