Poll 2018: Which Group of Pitchers Performs Better? — A Review

As I have at the end of the first half since 2013, I grouped two sets of pitchers together and aggregated their results through the half based on the degree of SIERA outperformance and underperformance. I then asked you which group of pitchers would perform better from an ERA perspective over the second half, and which range each group’s ERA would fall into. This year’s poll and voting results are here.

Let’s first review the poll results:

Which Group Posts a Lower 2nd Half ERA?
Group A – SIERA Outperformers 74.82%
Group B – SIERA Underperformers 19.07%
Neither, each group posts the same ERA, or within .05 in ERA 6.11%

Which Range Will Group A’s (the SIERA outperformance group) 2nd Half ERA Fall Into?
3.25-3.49 27.9%
3.50-3.74 25.72%
3.00-3.24 17.75%
3.75-3.99 15.94%
4.00-4.24 8.33%
2.75-2.99 1.81%
4.25 or above 1.45%
Below 2.75 1.09%

Which Range Will Group B’s (the SIERA underperformance group) 2nd Half ERA Fall Into?
4.00-4.24 34.69%
3.75-3.99 27.31%
4.25 or above 17.34%
3.50-3.74 12.55%
3.25-3.49 5.9%
3.00-3.24 1.48%
Below 2.75 0.37%
2.75-2.99 0.36%

By a vast majority, you believed that Group A, the SIERA outperformers, would post a lower second half ERA than Group B, the SIERA underperformers. Furthermore, more than half of you predicted that Group A’s ERA would land between 3.25 and 3.74 (their actual first half ERA was 2.83). On the other hand, more than half of you predicted that Group B’s ERA would settle in between 3.75 and 4.24 (their actual first half ERA was 5.49). So although you did believe in Group A’s ability to outperform their SIERA, you figured that the ERA gap between the two groups would narrow significantly.

Let’s find out how the two groups actually performed in the second half.

Ding ding ding, we have a winner! And it’s the nearly 75% of you who voted that Group A would post a lower second half ERA. But just about 2% of you rightly predicted that Group A’s ERA would finish in the 2.75 – 2.99 range. The Group B voters were far more prescient, as approximately 27% of you (representing the second most voted on ERA range) predicted the group’s ERA would settle into the 3.75 – 3.99 range.

So these results are both rather shocking and also as expected. Amazingly, Group A’s ERA declined from the first half, even after they posted a SIERA mark a whopping 1.35 runs greater in aggregate! Now granted, the second half ERA was just .02 runs lower, but maintaining a sub-3.00 ERA was not something I thought was possible. On the other hand, Group B’s performance did improve significantly, as they brought their ERA below 4.00, making them perfectly serviceable in even shallow mixed leagues. This after they torpedoed their team’s ratios in the first half.

So how did each group get to where they ended up?

Group A actually improved their skills sharply, as their strikeout rate spiked, thanks largely to Blake Snell and jumps from multiple pitchers. Their control also improved, as their walk rate dipped to exactly match Group B, which was one of the skills B had the advantage in. Though the group’s BABIP did rise, it still remained well below the league average and they even stranded a slightly higher rate of runners. Their HR/FB rate also increased, but remained below the league average. So the summary version is this — Group B got better and slightly less lucky, but the skills improvement offset the reduction in good fortune.

This list of outperformers includes many pitchers that are highly likely to be extremely overvalued in 2019 drafts. This is especially true of the low strikeout guys like Kyle Freeland (how on Earth did he do this while pitching half his games in the hitter’s haven known as Coors Field?!) and Miles Mikolas.

Group B also improved their strikeout and walk rates, but only marginally and not nearly to the same degree as Group A. These improvements dropped the group’s SIERA mark below 4.00. Surprisingly, now Group B actually outperformed their SIERA, though by a minor degree. Their BABIP declined precipitously, dropping from an inflated .323 to a better than league average .289. Do you really think these pitchers woke up in the second half with suddenly improved hits on balls in play suppression skills? No, of course not. It just goes to show how much a pitcher’s defense plays a role here. The lower BABIP allowed the group to strand a significantly higher rate of baserunners, though they posted a near identical HR/FB rate, which was higher than the league average.

***

It’s always important to remember that even one full pitcher season doesn’t represent enough of a sample size to conclude that a pitcher has BABIP or HR/FB rate suppressing abilities. While Group A’s continued SIERA outperformance might make you believe that the biggest outperformers do possess such skills and we shouldn’t necessarily bet against them, Group B reminds us how difficult these so-called luck metrics are to predict. If the 10 pitchers that underperformed their SIERA marks the most in the first half could end up posting an aggregate ERA below their SIERA in the second half, then we know luck plays a huge role in ERA. Focus on the underlying skills and you’ll be right far more often than wrong.