On Sunday we threw out some comps for Robert Gsellman, mentioning at the time that they were merely data points and that a full-blown comparison might show something else completely. So, what follows is an attempt to look for recent historical comps in a more vigorous manner. This is done with the understanding that Gsellman’s tiny sample is far less than ideal. But we can only work with what we have, not what we would find perfect.
Recall that Gsellman last year was in his age 22 season and in eight games – seven starts – he amassed 44.2 IP and posted a 2.42 ERA and a 2.63 FIP. So, a search was done from 1969, the first year of divisional play, to 2016 among rookie pitchers aged 20-24. Furthermore, the search was limited to guys who threw fewer than 100 innings while making at least five starts. Finally, it was for pitchers who recorded a FIP between 2.13 and 3.13, centering Gsellman’s mark (and age.) FIP was used hoping that even in the tiny sample it would be more predictive of future results.
There were 33 pitchers in our 48-year sample, showing how rare Gsellman’s numbers last year were. What follows is a chart of what the 31 pitchers did in their following season. Obviously, there are no numbers yet for Gsellman. And, surprisingly enough, there was a second pitcher who did this last year – the Cardinals’ Alex Reyes. The players below are listed in order of their initial FIP.
| Player | GS | IP | ERA | Notes |
|---|---|---|---|---|
| Pat Combs | 31 | 182 | 4.05 | Walk rate tripled |
| Bill Slayback | 0 | 0 | 0.00 | Arm problems |
| Andy Rincon | 6 | 31.2 | 5.40 | Line drive to arm |
| J.R. Richard | 9 | 50.2 | 4.44 | 31 BB |
| Don Durham | 4 | 11.2 | 13.11 | 2.743 WHIP |
| Nick Tropeano | 13 | 68.1 | 3.56 | 15.1 HR/FB |
| John Martin | 15 | 97.1 | 3.42 | Numbers held down by ’81 strike |
| Brandon Cumpton | 10 | 56 | 4.98 | 3.02 FIP |
| Alex Wood | 24 | 156.1 | 2.59 | 1.094 WHIP |
| Scott Sanderson | 24 | 144 | 3.56 | Pitched a CGSO against the Mets in his penultimate start |
| Sonny Gray | 33 | 219 | 3.08 | 14-game winner |
| Pedro Astacio | 31 | 186.1 | 3.57 | 14-game winner |
| Marty Bystrom | 9 | 53.2 | 3.35 | Limited by shoulder problems |
| Jim Colborn | 0 | 0 | 0.00 | Injury? Won 20 in ’73 |
| Joel Pineiro | 28 | 176.1 | 3.52 | Won 13 as a SP |
| Wade Davis | 29 | 168 | 4.07 | 4th in ROY balloting |
| Michael Wacha | 19 | 107 | 3.20 | Shoulder blade injury |
| Dave Pagan | 0 | 0 | 0.00 | Injury? Only 62 IP majors & minors |
| Zach Duke | 34 | 215.1 | 4.47 | Led league in hits allowed 255 |
| Dave Lemonds | 0 | 0 | 0.00 | Injury? Only 7 games in minors |
| John Butcher | 3 | 21.2 | 2.08 | Spent most of year in minors |
| Nino Espinosa | 29 | 195.1 | 3.50 | Does anyone not like this guy? |
| Kevin Millwood | 29 | 171.2 | 4.14 | 17-game winner |
| Randy Tomlin | 27 | 169 | 3.09 | No relation to Josh |
| Jon Matlack | 32 | 238 | 2.34 | Rookie of the Year |
| Dan Larson | 10 | 61.2 | 6.42 | Former 1st-round pick traded for Claude Osteen |
| Dave Eiland | 13 | 62.1 | 5.20 | 20 BB and 16 Ks as SP |
| Tyler Thornburg | 0 | 0 | 0.00 | 27 games as a reliever |
| Trevor Bell | 0 | 0 | 0.00 | 19 games as a reliever |
| Hung-Chih Kuo | 6 | 28 | 6.75 | Made six starts in June |
| Rich Wortham | 33 | 204 | 4.90 | 14-game winner faded in September |
Gsellman had the ninth-best FIP in this group while Reyes was 11th. What’s nice about this group is the wide mixture of names included. There are top draft picks and there are guys who were complete afterthoughts. There are guys who went on to have great careers, guys whose careers were cut short by injury and guys who simply never amounted to much.
Our primary interest is to see what they did in the following season, specifically as a starting pitcher. The stats listed above are only what they did as starters. It’s more difficult to come in and put up good numbers multiple times through the lineup. What Matlack did as a starter is so much more impressive than what any of these guys might have done with bullpen-aided stats, so those were eliminated.
Of our 31-player sample, seven can be eliminated completely, as they did not start a single game the next season. Then you just have to draw the line on where is a good number of starts for a youngster breaking in the majors. Let’s start off using 19, which includes Wacha in our group. He was clearly a SP and didn’t get more starts simply because of injury.
That gives us 14 pitchers.
Of these 14, nine posted an ERA under 4.00 or more specifically between 2.34 and 3.57 which is pretty darn good. Three more posted an ERA in the low 4s (4.05, 4.07 and 4.14) and two had elevated ERAs. Of course, there’s a survivor bias inherent here. If a guy stinks, his team isn’t going to allow him to get 19 starts. Kuo got six consecutive starts and if he had been good, he would have gotten more. But he was terrible.
But guys didn’t reach this mark for other reasons, too. Bystrom certainly would have gotten 19 if he was healthy. Martin would have if it wasn’t for the 1981 strike. It’s depressing to see how many of these guys ran into arm injuries. Yet another reminder of how fragile pitching is. Seven of these 31 guys likely had arm injuries and possibly others, too.
If you’re looking for a definitive answer, you’re not going to find it here. If you want to be bullish and name Matlack, Gray and Pineiro someone else can be bearish and name Larson, Eiland and Kuo. Here’s the best we can do: We started with 31 players and eliminated the seven who didn’t have a single start the following season, giving us 24. Of those 24, 16 had an ERA as a starter the following year of 4.14 or less, with the majority being 3.57 or less. Of those 16, one (Butcher) had just three starts. But four of the eight with high ERAs had nine or fewer starts. And one of those was Richard, who ended up having an All-Star career.
With a bigger overall sample, we could have looked to focus the study group even further. Does Gsellman’s 13th-round draft status make it more likely for him to fail? Does his high-strikeout, high-grounder results make it more likely for him to succeed? But our sample simply wasn’t big enough to make these additional distinctions.
My preference going into this was for Gsellman to get the opportunity to be a starter in 2017. While the data wasn’t conclusive in either direction, in my opinion it seemed to be more bullish than not. Still, we are going off a small initial sample of 44.2 IP and our comparison group was not as robust as we would have preferred.
He had movement and usable pitches and good velocity and deception. He certainly looks the part of a guy who can have a career.
I look at his MILB stats and see that he had a very nice year at Bingo…his first 3 starts for Vegas were a bit of a Bombfest—- his following 6 starts for Vegas were more in line with his total year. We never really know what a Pitcher is working on at the minor league level, but I’d make an assumption that the Stats and production are better “project-ables” at AAA versus the lower levels.
As a general rule, sure. But we can’t ignore that Las Vegas is a hitter-friendly park in a hitter-friendly league.
But in his last six starts for Las Vegas, we see a lot of the same traits we saw in the majors. He had 35 Ks and 12 BB, a 2.92 K/BB ratio. In the majors he had a 2.80 rate.
He had 35 Ks in 38.2 IP in those last six starts for an 8.1 K/9 rate. In the majors he had an 8.5 rate.
For walks it was 2.8 in the minors and 3.0 in the majors
The big difference is the homers. In his last six minor league games, he allowed 6 HR in 38.2 IP and in the majors it was 1 in 44.2 IP. We know the major league HR rate is going up next year. But I would expect it to be less than his Triple-A rate.
Interesting assessment. I fully get the rearward looking comps to place his contribution in “context”, but I cant see how any player from any time past will predict the outcome of a different player at a different time, playing under different conditions. In any event, I think we saw enough of Gsellman to get him to ST and let him compete for a starting pitcher role. If he fails, then let him earn it in LV.
I was hoping that extending the time frame back to 1969 would have produced a larger sample, a minimum of triple digits. But that didn’t happen. Pulling numbers out of thin air — if we had gotten 200 people in the comparison group and 70% of them had posted an ERA 3.75 or lower, I would have felt very good about it all. I would have preferred to use ERA+ to eliminate the era differences but that became too much when I eliminated the relief appearances. Zach Duke’s 100 ERA+ looks a lot better than his 4.47 ERA.
If its any value, I argue with my students regularly about data set size and statistical v. deterministic outcomes!
Im in awe about the hard work you put into provide these kinds of exercises. Very cool!
Thanks for the kind words!
And just to make it clear for others who may be reading – nothing comes with a guarantee, even if we had a fabulously big sample size. All we can do is play the percentages and hope that over the long haul that leads us to more right assumptions than wrong ones.
Chris, I second that. As I was reading it I was asking myself if Brian was on vacation and had this much time but if not, where does he find it?
I barely get a chance to read some of the articles, much less research it this deeply.
He had 158 innings in Bingo in 2016 and 2016 combined— 6 Homers. The Balance of MILB numbers tell you that he was either too far above the competition, or that he’s been Hard to Lift.
Brian,
There are times when your statistical explanations make me think you are related to Nate Silver of 538 fame. I mean that as very high compliment.
Familia arrested on domestic violence charges.
Well, its gonna be closer shopping…
Addison – I know your aim is true.
Excellent work on the comp analysis. Despite the small sample, Gsellman passed the eye test IMHO. Additionally, he pitched in high stress games and looked composed beyond his years. Despite the uncertain projections, both he and Lugo look to lie good assets that can contribute at the MLB level in 2017. I would be reluctant to deal either unless they are part of a deal that returns a high value every day player.
I am also bullish on Gsellman as a starter. It is very easy to get excited because of his peripherals. Your excellent analysis tempers the excitement and sheds light on the situation. If the Mets have five healthy starters in Syndergaard, deGrom, Harvey, Matz and Wheeler then where would Gsellman fit in? Six man rotations don’t seem to be popular among starting pitchers. If the Mets would skip each starter every six times through the rotation and the designated spot starter would take his place, then only the spot starter would be out of sync with a five man rotation. It would work out to 26 starts each for the six pitchers, with an average of 156 innings pitched (at the major league average 6 innings per start), thus having a well rested staff for the post season. Gsellman would be that sixth spot starter.