For a Starter to Beat His ERA Estimators …

The “ability” of a pitcher to consistently beat his ERA estimators will always be a discussion top. Today, I’m going to put context on who has suppressed their ERA for two straight seasons and how they performed in the third season. I’ve been trying to see if I have missed anything while digging into under and overperforming starts and found that I might have missed the obvious, the starter’s team.

Before getting to the team context, here are the baseline chances for starting pitchers to consistently beat certain ERA benchmarks.

Beat ERA Benchmarks

The requirements I used to select the pitchers in question were:

• Seasons since 2005.
• Over half of the pitcher’s appearances must be starts.
• Minimum 80 IP in each season except 2020 where the minimum was 40 IP

For the first test, I examined pitchers who beat their SIERA by 0.50 in the first two seasons and then how they performed in the third season. Here are the results (count = 120).

A few notes from the table.

• The average difference dropped from about 0.90 down to 0.20. The pitchers still beat their estimators but by not as much.
• The pitchers are still posting a decent BABIP in the third season with 47 (39%) of them still outperforming their BABIP by 0.5 runs. Looking that the rest of the year three results, 21% beat the projections by less than half a run, 24% had their ERA underperform by less than half a run, and finally, and for the final 16%, their ERA was over a half run higher than the estimator. The key takeaway for me would be that 40% of the starters who beat their ERA estimator for two straight seasons can’t beat the ERA estimator in the third year.
• I looked into several variables and the key factor common in pitchers who beat the estimator was a low BABIP. Some may have also limited home runs or posted a low left on-base percentage, but the non-BABIP values evened out over the whole sample.
• Not one batted ball-generating pitcher (e.g. groundball) stood out. The ERA estimators, especially SIERA would already take these into account.

Next, I upped the threshold to 0.75 runs. I wanted to go higher be the sample size was getting too small. Here are the number of pitchers who meet each difference criteria (the “1” was Zach Davies starting in 2019).

And here are the results for a .75 run difference.

In this case, the original differences averaged out to be well over a run but in the third season, the difference was just 0.12.

Here are the percentages that the pitcher had a certain ERA difference in season three.

ERA-SIERA: Chance
>0.75: 22%
0.00 to 0.75: 37%
-0.75 to 0.00: 29%
< -0.75: 12%

With this sample, 59% maintain some outperformance, almost identical to the 0.50 sample.

Team Context

While the core reason for the low ERA was BABIP related, the reason for the low BABIP might not be pitcher-related but team related.

First, here is the percentage chance that the 0.50 pitchers were on the same team.

And for the 0.75 Sample.

The 0.50 sample results are cleaner (amazing what a larger sample size will do) but both point to pitchers who remained with the same team keeping more of the advantage. Here is how the pitchers who remained with the same team performed in the third season compared to those who changed teams.

So, on average, any starter who kept their some of their gains remained with the same team while those who changed teams had their ERA regress back to their SIERA. Determining who might have this advantage is the next step but it’s not an easy one. Some same-team pitchers will get the advantage because the team’s strength (e.g. great infield defense) matches up with the pitcher’s strength (e.g. generating groundballs). Also, the catcher or team’s attack plan could be the key factor. At this point, it is tough to determine the exact cause so I’ll call it a day. I’ll mull over the results for a while and determine my next course of action.