Should Projections for Astros Hitters be Tempered?
Yesterday, I noticed someone slid into my DM’s and I got my hopes up but instead I got this:
Maybe. Many of the Astros players didn’t think it helped.
Some Astros players told my investigators that they did not believe the sign-stealing scheme was effective, and it was more distracting than useful to hitters.
We just don’t know for sure of the effects of cheating so I guess I better take a stab and find out.
To start with, I went to the projection sources to find out how the projections weigh each year’s results. The weighted averages, along with some aging adjustments and regression, create the final projections. ZiPS is up first.
Dan Szymborski uses individual weightings for each component (strikeouts, doubles, etc) but at the end, the weighting is close to 8-5-4-2 where ‘8’ is the last season. According to the commissioner’s report, the Astros “only” cheated at home in 2017 so only 2 units (half of four, the third value) of the weighting will be boosted. The percentage of the projection’s input from the cheating is 10.5% (2/[8+5+4+2] or 2/19).
Assume for now the Astros hitters received a 10% increase in production by cheating. A hitter’s 2020 projection will be inflated by just 1% (10% of 10.5%).
Jared Cross informed me the Steamer projection weights are close to 5.0, 3.2, 2.0, 1.3, 0.8. If still assuming the same 10% more production from cheating, the projection boost percentage is 0.8% or less than 1%
(2.0/2)/(5.0+3.2+3.0+1.3+.8) * .10
1/12.3*.10
.081*.10
0.008
While I used 10% in the above examples, there are a couple of ways to see how much the trash can banging helped. The first and easiest method is to compare how the Astros did at home and away during the years in question. With wRC+ on a 100 scale where 100 is the league average, it’s easy to measure the percentage change. Here are the Astros wRC+’s for the past four seasons with the yearly change.
| Season | Home wRC+ | Difference | Away wRC+ | Difference |
|---|---|---|---|---|
| 2016 | 100 | 103 | ||
| 2017 | 121 | 21 | 122 | 19 |
| 2018 | 107 | -14 | 113 | -9 |
| 2019 | 136 | 29 | 116 | 3 |
From 2016 to 2017, the Astros had a gain of 2% points higher at home than on the road. Then they gave up all the gains and more from 2017 to 2018. The real anomaly happened from 2018 to 2109 where the home wRC+ jumped from 29% points while the value on the road was just 12% points.
If the small difference from 2017 to 2018 is going to be used, the huge disparity this past season must then be explained. According to MLB’s report, the Astros didn’t admit or get caught cheating in 2019.
Let’s ASSUME for second, the 26% difference in production was from cheating, and the Astros lose the advantage. The weighting would be more because the results just occurred and the home weighting is 21%. So 26% of 21% is a 5% drop. That’s a bigger deal. I know … 5% > 1%.
Next, I found out how common is a 21-point difference between the home and away values (2016 to 2019). It was the 7th highest difference with the 2016 Indians at 29 (115 wRC+ at home, 86 wRC+ on the road) being the most extreme. The difference is high but not unreasonable.
The 29-point jump from one season to the next was the 3rd largest jump with the 2016 to 2017 Athletics increasing from 80 wwRC+ to 117 wRC+. The Twins road numbers this past season was the second biggest jump from 88 wRC+ to 121 wRC+. Yes, the production jump at home was significant but just not unexplainable.
One issue with just using the overall team numbers is that the composition of the team could change from season to season. My next test takes the preseason steamer OPS projection (no wRC+ in the dataset) and sees how the hitters’ actual production compares. I’ve weighted the OPS values based on the players’ plate appearances. Also, I included the league-wide OPS values since offense production has been increasing the past few seasons.
| Season | Change in Home OPS | Change in Road OPS | League Change | |
|---|---|---|---|---|
| 2017 | .055 | .077 | 16 to 17 | 0.012 |
| 2018 | -.055 | -.005 | 17 to 18 | -0.022 |
| 2019 | .109 | .066 | 18 to 19 | 0.029 |
| Total | .109 | .138 | 16 to 19 | 0.019 |
The table closely mirrors the previous table with equal jumps in 2017, drops in 2018, and back up in 2019. The most interesting part to me is that the Astros hitters over the three seasons improved more on the road than at home.
With the above facts, I think the projections could be inflated from cheating by 1% on the extreme end. Most likely it’s zero. Zip. None. Since the banging happened three years ago, the “boost” has been watered down. Just feel free to draft away those cheating bastards with no reservations of the hitters missing their projections.
While some beat your favorite author leagues will be starting soon, feel free to go to the NFBC and sign up for their Draft or Online Championships. Also, for those with deeper pocketbooks, the NFBC Main Events are quickly filling up.
Jeff, one of the authors of the fantasy baseball guide,The Process, writes for RotoGraphs, The Hardball Times, Rotowire, Baseball America, and BaseballHQ. He has been nominated for two SABR Analytics Research Award for Contemporary Analysis and won it in 2013 in tandem with Bill Petti. He has won four FSWA Awards including on for his Mining the News series. He's won Tout Wars three times, LABR twice, and got his first NFBC Main Event win in 2021. Follow him on Twitter @jeffwzimmerman.
I saw a video of the cheating in action – Danny Farquhar figured out the Astros were sign-stealing in one game and changed up his signs. Somebody went and found video of the PA and it was against Evan Gattis in 2017 in a game that was fairly sparsely attended so you can hear the lid-banging in the background.
(BTW, video is NSFW for a swear word or two (or four) in the embedded video, not in the story itself but in the video)
https://ftw.usatoday.com/2019/11/astros-fiers-stealing-signs-danny-farquhar-video-mlb
I was initially going to question how useful that really was, but re-watching it I think Gattis gets the signal early enough that it could help. To me the real issue would be the fastball where you get no lid-banging and you’re waiting for the lack of a signal. Seems like it would take you very close to when the pitch is actually delivered and it’s awfully late to adjust at that point. I will say, seems to me like a huge advantage with runners on where the pitcher has to wait a split-second longer.
Entirely right. I get the outrage but but the significance is absolutely overstated. I work with a guy who now wants to pretend that Dodgers have a few WS rings lol.
AGain, the Astros said it wasn’t useful in 2019 because they couldn’t figure out the sign rotation quick enough…i.e, so many teams knew they were doing it (and others – the Astros are NOT the only ones) that catcher to pitcher signs are now a rotating algorithm almost. It’s crazy. But, it probably didn’t help them last year although god forbid they do figure it out and you don’t change it from start to start, then you get bushwacked like Glasnow did in Gm 5. He said he was tipping, but I’m betting he didn’t change his signs from the last game and the Stros figured it out by reviewing the last game. as soon as they saw the same rotation they jumped on it. He changed it like 5 batter in but by then it was too late.
Rays fans still making excuses. Tipping pitches is not sign stealing. I have watched that series of ABs many times. They figure it out on the 3rd batter of the inning and it gets changed two hitters later. At most two hitters benefited. Glasnow got beat that day.
Any thoughts on whether there would be any ramifications from the upheaval? New manager, GM, added press, more scrutiny, etc?
Great analysis. But to really see if there was a true difference, I don’t think you should look at the whole team. It could be argued that knowing the pitch coming in might only be advantageous to an elite hitter, rather than just an average one. For the same reasons an elite hitter is elite, giving them an advantage (knowing the pitch) would result in a greater delta than if they were simply average and didn’t have the talent to take advantage of that knowledge. Though exaggerative, I can be tipped off to a fastball all day, I’m never hitting that ball. Same for an average hitter, who even when guessing right can’t hit the heater. An elite hitter however, when you take away the need to “guess” on a pitch removes any opportunity for the pitcher to fool the hitter (especially for off-speed pitchers).
My recommendation is to do exactly what you did, but only for the top of the lineup, or high-WAR guys. I’d be willing to bet that you’d see a better divergence once you drill it down.
Big difference when an elite hitter says “Thank god i guessed right on that fastball” instead of “heck YEAH I sat back on that fastball!”
Every real discussion I have ever heard is that some hitters want to know what is coming and some don’t. Its not about ability but mentality. Being wrong if you “know” what is coming is worse than a normal AB, which they make a living from doing well. As for your last sentence, I think it is the same sentence twice. One is a fictional mentality and the other is someone with some confidence, which is every professional athlete. Very basic hitting is sitting on FB and adjusting to offspeed – nobody has to guess fastball in MLB unless it is coming 105.
I would not expect the divergence to have anything to do with lineup position, WAR or anything like that. It would be select guys that wanted the information. I would expect them to be bad hitters actually.
Fair point. And to you point, if some hitters are gaining advantage, and other are not, then it would be reasonable, from a statistical side of things, to not look at the team in totality, but look at the individual deltas home and way, and see if some benefited in a tangible way or not. My point (and I think yours too) is that if all players are not benefited (or hurt) in a uniform direction, then you can’t look at an average of all. Because the effects of the benefit on certain batters will be diluted or negated by the negative effects on certain batters.
And let’s be realistic here. The Dodgers didn’t lose world series’ because the 6 to 9 batters were crushing the ball. They were losing because the high WAR/OPS guys were murdering their aces’ pitches, from Kershaw to Darvish.
Just do the analysis and let’s see what it turns up.
I like this article, a lot. Good logic. Astros claim 8 other teams did similar schemes. I’d even say the benefit may have been a wash if all the top teams were doing it. We know the Yankees and Red Sox had a teenage fight about it in 2017 blaming one another. Might have been a battle of all the cheaters cheating and getting mad at the other after they won regardless if cheating helped them.
LOL all the top teams doing it. I don’t doubt that it was widespread. There is no reason it would be confined to top teams. I bet you that if the Astros won the WS this story would not have even come out as it wouldn’t have been profitable for MLB. No reason to pick on teams that already have problems as that isn’t profitable either. Unlikable teams are good for business and they have already profited plenty. Is the timing not interesting? I think it is convenient that this “news” cancelled out the juiced ball scandal that was building. Its almost like MLB traded an organizational problem for a few problems confined to very well off individuals. Can you think of better scapegoats? Well, I am sure we could drag Barry Bonds into the mix but that might be stretch.
Last year, I picked up Gurriel mid-season, while he was having a power surge. I remember distinctly being wary that the surge was somewhat strange and watching a few at bats closely, one of which he jacked a homer out on. On that AB, I remember distinctly thinking that he loaded up on a pitch, in a way that suggested he knew what was coming.
You should have picked him up in 2017, no?
No. I am sure the significance of this is being overstated. This is news after all and the year is 2020. In reality I bet a few players were really into it, just guessing Bregman based on a few things I have seen, but most probably didn’t benefit much at all. There seems to be a lot of data that it didn’t make the difference that people would like it to have made. Stats change year over year and there is never a scientific explanation for it. Its not a die that you are rolling, they are humans using round things vs other humans with round things – there will be variance. I can’t wait to start talking about baseball again at some point. Once this blows over we will get back to juiced ball manipulation. I have a sick feeling I might be waiting a long time!
Seems to me any advantage would be in knowing when an off-speed pitch is coming, so that you can exclude the fast ball in your approach. I agree with commenter that might also apply only to elite hitters at hitting off-speed pitches. Seems an analysis could be done comparing elite Astro hitters during this time even home/road to success on off-speed pitchers. Then throw out pitchers that are junk ballers anyway, or have no credible off speed pitch, and maybe even cone down further on pitchers with known plus FB and offspeed. Then figure that it is unlikely this technique is used all the time – just in key situations. Then figure that even if you are an elite hitter and know an offspeed pitch is coming, you still only have at best a 1/3 chance of hitting it – but the question is whether that is 3/10 or 4/10.
I don’t think it makes a difference in projecting 2020 stats. But Cora, Belttan, and some Astro hitters thought it helped win games.