Advertisement

Pitching by the Numbers: Luckiest and unluckiest fantasy starters

Michael Salfino
Pitching by the Numbers: Luckiest and unluckiest fantasy starters

We’re told that batting average allowed is largely luck, making so much bottom-line pitching performance (ERA and WHIP) largely random. This is depressing. But the solution isn’t to curse that our faults lie in the stars but rather to strive to better isolate luck by cross-checking batting average against contact type.

We use well-hit data here from Inside Edge, where scouts review each batted ball for whether, to their eyes, it was well struck. Balls out of play — homers — are counted as they should be as well-hit. Grounders are assessed. Not all line drives are well hit. Strikeouts count because the stat is tethered to at bats. The MLB well-hit average this year is .134, meaning that pitchers allow batters to hit the ball well 13.4 percent of the time.

[Join a Yahoo Daily Fantasy Baseball contest now]

We also know that, as of action through Wednesday, the MLB-wide batting average was .251. So a pitcher’s batting average allowed is expected to be .117 higher than his well-hit rate. To make the math really simple, what I do on the chart this week is to simply add .117 to each pitcher’s Inside Edge well-hit rate in order to calculate an expected batting average. Then I compare this expected batting average to the actual batting average each pitcher allowed. We’re highlighting only the most fantasy-relevant outliers, meaning the pitchers who, according to well-hit average, should have a batting average much higher or lower than their actual one.

MLB Pitchers with Greatest Difference in WHA and BAA | PointAfter

Now of course, we’d all like to reject the randomness of baseball statistics and believe that they are a product of pure skill. One argument for this is that the high variance itself proves that there is skill involved in the difference between well-hit rate and average. But this makes no sense because if we stipulated that batting average allowed was completely random we would not expect every pitcher to be exactly the same in the statistic (even if they played exactly the same teams).

And there of course can be other factors here that would dissuade one from betting on regression to the norms. For example, a pitcher can be heavily ground ball or fly ball or have defenses that are above or below average or have teams that more effectively use defensive positioning/shifts. But we’re not going to overfit by looking to perfectly explain what happened. We’re just trying to better predict what will happen going forward in one simple way each week.

[Yahoo Fantasy Football is open for the 2016 season. Sign up now]

Matt Harvey is the unluckiest pitcher in baseball by this metric. His well-hit rate is below average —very unusual for him because it’s always been great —but the league should be hitting .262 against him, not .330. This doesn’t mean to run out and buy Harvey as much as it means that Harvey’s bad season has a large luck component — maybe half bad luck and half poor performance. So he’s closer to being good than his ERA and WHIP indicate.

The same can be said for Michael Pineda and Adam Wainwright, though again they are about average in well-hit average allowed. Dallas Keuchel is a buy but I don’t see enough of a discount. The better buy relative to cost according to this stat is Carlos Rodon, who actually has a pretty sterling well-hit rate and has also progressed greatly in BB percentage. 

Noah Syndergaard has been very unlucky but again the Mets’ defense is admittedly subpar and they shift less than any team in baseball. Still it’s nice to know that Syndergaard is this good without any luck and even, as far as we can tell, a lot of bad luck.

A struggling starter that this model says NOT to bet on is Chris Archer, who is just getting pounded in well-hit and who should have an average allowed 41 points higher. Again, though, the Rays have generally been a good BABIP team so maybe this is sustainable. But is Archer likely to get better given his present well-hit rate? No. He’s earned all of his bad stats and then some.

You can see that Drew Pomeranz is for real (for as long as he can sustain his workload). Jason Hammel and John Lackey are borderline. Do not buy them but I can’t see getting that much for either in a trade because I don’t sense there are a lot of believers out there. 

Justin Verlander has been very lucky and has brand value still. I’m not going to say to trade Chris Sale or Jake Arrieta despite their variance between expected and actual average because both are elite in well-hit rate. Jose Fernandez pitched very well the day after this chart was submitted so let’s not do anything with him right now either. 

Jose Quintana is the biggest sell on the list. That is a terrible well-hit rate and makes his batting average allowed almost miraculous. But Sale is on the lucky list too so maybe the White Sox are the anti-Mets in having a very contact-proof defense. This is where FIP is a bad stat since his BABIP says his ERA should be 2.08, lower than his actual 2.22 — cite this if you trade him as I advise. I would bet well-hit far more heavily than straight FIP/BABIP because well-hit is a far more specific and descriptive stat.