SoxProspects News
|
|
|
|
Legal
Forum Ground Rules
The views expressed by the members of this Forum do not necessarily reflect the views of SoxProspects, LLC.
© 2003-2024 SoxProspects, LLC
|
|
|
|
|
Forum Home | Search | My Profile | Messages | Members | Help |
Welcome Guest. Please Login or Register.
|
Post by philsbosoxfan on Oct 10, 2020 12:18:23 GMT -5
How good were the hitters Houck faced? I grabbed every hitter's season performance against each of his three pitches, and adjusted them for what the whole league did. (The league had a wFA/C of 0.08, a wSI/C of -.02, and a wSL/C of -.53. Which is to say, 4-seamers were hit .08 runs per 100 pitchers better than average, but sliders were hit -.53 runs below average. In a separate study of pitchers 2015-2020 I came up with -.54 for the latter figure, so these are likely very reliable).
Houck's hitters were: +1.08 runs/100 pitches against the FA (he threw 98)
+1.65 against the sinker (he threw 65)
+0.16 against the slider (he threw 94)
Overall, they were 0.92 against the three pitches.
The sinker number helps explain why he didn't throw it more than than the 4-seamer. Players who saw 120 sinkers and had a 1.50 to 1.80 against them: Acuna, Betts, Matt Carpenter, Bellinger, Freeman, JBJ, Bregman.
Houck had a 1.85 wFA/C, but that was really a 2.93 when you factor in his hitters. He had a 1.99 wSI/C, but that was 3.64 against average hitters. The slider goes from 4.12 to 4.28. Now, these include his BABIP luck. But it gives you a better idea of the relative effectiveness of his pitches.
I'm going to try to figure out how to take out the BABIP luck next. Whether or not I can do that, I want to get a sense of where he would rank in MLB on the three pitches.
Separate question. If the implication of the lack of sinkers were a factor of planning rather than bullets in the holster on any given day, iirc, most of his sinkers were against the Yankees. Wouldn't that imply that the Yankees were relatively weak against sinkers compared to the Marlins and Braves ? Rephrasing that to what I really am curious about. During the Yankee game, it was assumed that he went to the sinker because he didn't have his 4 seamer going well. If instead, that was the game plan, that's a different animal altogether.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,931
|
Post by ericmvan on Oct 10, 2020 18:16:01 GMT -5
How good were the hitters Houck faced? I grabbed every hitter's season performance against each of his three pitches, and adjusted them for what the whole league did. (The league had a wFA/C of 0.08, a wSI/C of -.02, and a wSL/C of -.53. Which is to say, 4-seamers were hit .08 runs per 100 pitchers better than average, but sliders were hit -.53 runs below average. In a separate study of pitchers 2015-2020 I came up with -.54 for the latter figure, so these are likely very reliable).
Houck's hitters were: +1.08 runs/100 pitches against the FA (he threw 98)
+1.65 +0.70 against the sinker (he threw 65)
+0.16 against the slider (he threw 94)
Overall, they were 0.92 0.68 against the three pitches.
The sinker number helps explain why he didn't throw it more than than the 4-seamer. Players who saw 120 sinkers and had a 1.50 to 1.80 against them: Acuna, Betts, Matt Carpenter, Bellinger, Freeman, JBJ, Bregman.
Houck had a 1.85 wFA/C, but that was really a 2.93 when you factor in his hitters. He had a 1.99 wSI/C, but that was 3.64 2.69 against average hitters. The slider goes from 4.12 to 4.28. Now, these include his BABIP luck. But it gives you a better idea of the relative effectiveness of his pitches.
I'm going to try to figure out how to take out the BABIP luck next. Whether or not I can do that, I want to get a sense of where he would rank in MLB on the three pitches.
Separate question. If the implication of the lack of sinkers were a factor of planning rather than bullets in the holster on any given day, iirc, most of his sinkers were against the Yankees. Wouldn't that imply that the Yankees were relatively weak against sinkers compared to the Marlins and Braves ? Rephrasing that to what I really am curious about. During the Yankee game, it was assumed that he went to the sinker because he didn't have his 4 seamer going well. If instead, that was the game plan, that's a different animal altogether. Great question! It generated lots of interesting data. But first, the sinker numbers were wrong. I've corrected them above. Opponent lineup strength against each of his three main pitches, and how many he threw. The pitches are FA, SI, SL, and the number is just the average of the 9 hitters.
Miami was .13, 1.14, -0.66. He threw 35, 15, and 30. NYY was 1.16, -0.32, -0.93. He threw 20, 36, and 37.
So, yes, unlike Miami which hit the sinker much better thgan the 4-seamer, the Yankees were the opposite and this was reflected in their use. I'm guessing it was partly the plan and partly what was working. However:
Atl was 2.19, 0.91, and 0.10 He threw 43, 14, and 37. Pitching to FB strength, not weakness!
So, for each of his 27 hitters I calculated the correlation between how much difficulty a guy had with each pitch, and how many of that pitch they threw him. Those results are so interesting I post them in full. A 1.000 means they attacked the hitters precisely as expected, while a -1.000 means they completely pitched to his strength. Name Tm Attack Gary Sanchez NY .990 Clint Frazier NY .988 Gianca. Stanton NY .945 Travis d'Arnaud Atl .899 Aaron Hicks NY .754 Matt Joyce Mia .682 Corey Dickerson Mia .614 Jazz Chisholm Mia .491 Luke Voit NY .446 Tyler Wade NY .354 Miguel Rojas Mia .294 Marcell Ozuna Atl .286 Adam Duvall Atl .216 Starling Marte Mia .186 Ad. Hechavarria Atl .126 DJ LeMahieu NY .101 Gleyber Torres NY .095 Ender Inciarte Atl .088 Jesus Aguilar Mia -.329 Dansby Swanson Atl -.450 Jorge Alfaro Mia -.509 Freddie Freeman Atl -.837 Ozzie Albies Atl -.896 Garrett Cooper Mia -.950 Brian Anderson Mia -.955 Brett Gardner NY -1.000
There are 8 guys between 0 and .3 and nobody between 0 and -.3. That really suggests that the last 8 guys were consciously pitched backwards.
Obviously, I'll take a look at some or all of those guys later.
Meanwhile, I have a bunch of really interesting stuff about BABIP karma as measured by wOBA - xwOBA, but the teaser trailer is that it looks like he has a 50% chance of having a skill.
|
|
|
Post by philsbosoxfan on Oct 10, 2020 18:52:42 GMT -5
Separate question. If the implication of the lack of sinkers were a factor of planning rather than bullets in the holster on any given day, iirc, most of his sinkers were against the Yankees. Wouldn't that imply that the Yankees were relatively weak against sinkers compared to the Marlins and Braves ? Rephrasing that to what I really am curious about. During the Yankee game, it was assumed that he went to the sinker because he didn't have his 4 seamer going well. If instead, that was the game plan, that's a different animal altogether. Great question! It generated lots of interesting data. But first, the sinker numbers were wrong. I've corrected them above. Opponent lineup strength against each of his three main pitches, and how many he threw. The pitches are FA, SI, SL, and the number is just the average of the 9 hitters.
Miami was .13, 1.14, -0.66. He threw 35, 15, and 30. NYY was 1.16, -0.32, -0.93. He threw 20, 36, and 37.
So, yes, unlike Miami which hit the sinker much better thgan the 4-seamer, the Yankees were the opposite and this was reflected in their use. I'm guessing it was partly the plan and partly what was working. However:
Atl was 2.19, 0.91, and 0.10 He threw 43, 14, and 37. Pitching to FB strength, not weakness!
So, for each of his 27 hitters I calculated the correlation between how much difficulty a guy had with each pitch, and how many of that pitch they threw him. Those results are so interesting I post them in full. A 1.000 means they attacked the hitters precisely as expected, while a -1.000 means they completely pitched to his strength. Name Tm Attack Gary Sanchez NY .990 Clint Frazier NY .988 Gianca. Stanton NY .945 Travis d'Arnaud Atl .899 Aaron Hicks NY .754 Matt Joyce Mia .682 Corey Dickerson Mia .614 Jazz Chisholm Mia .491 Luke Voit NY .446 Tyler Wade NY .354 Miguel Rojas Mia .294 Marcell Ozuna Atl .286 Adam Duvall Atl .216 Starling Marte Mia .186 Ad. Hechavarria Atl .126 DJ LeMahieu NY .101 Gleyber Torres NY .095 Ender Inciarte Atl .088 Jesus Aguilar Mia -.329 Dansby Swanson Atl -.450 Jorge Alfaro Mia -.509 Freddie Freeman Atl -.837 Ozzie Albies Atl -.896 Garrett Cooper Mia -.950 Brian Anderson Mia -.955 Brett Gardner NY -1.000
There are 8 guys between 0 and .3 and nobody between 0 and -.3. That really suggests that the last 8 guys were consciously pitched backwards. Obviously, I'll take a look at some or all of those guys later. Meanwhile, I have a bunch of really interesting stuff about BABIP karma as measured by wOBA - xwOBA, but the teaser trailer is that it looks like he has a 50% chance of having a skill.
Good stuff. I didn't ask about specific players mainly because there were some batters he was clearly staying away from which is a different approach altogether, Freeman for example, stands out. It would be pretty subjective (and a lot of work) to single those out. Since he went opposite for Atlanta, I'd assume it wasn't a general game plan but more likely a player specific plan with the above issue of attacking or not attacking specific batters.
|
|
|
Post by philsbosoxfan on Oct 12, 2020 4:34:09 GMT -5
Thinking about it, it's amazing that now with the data available, you could hypothetically rank each pitchers pitch vs the quality of the competition he faced, do the same for hitters then re-run the data using adjusted numbers for both. Hypothetically a game plan for every pitcher and hitter against every hitter and pitcher.
If you added in factor like location, velocity and break, the possibilities are almost endless. Hypothetically player A hits big break curves well but struggles against small break curves, etc. etc. etc. Other factors, bats R/L throws R/L.
That all said, I'm looking forward to Houck's adjusted slider ranking numbers to see how well it correlates to the break ranking numbers.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,931
|
Post by ericmvan on Oct 13, 2020 12:38:34 GMT -5
Thinking about it, it's amazing that now with the data available, you could hypothetically rank each pitchers pitch vs the quality of the competition he faced, do the same for hitters then re-run the data using adjusted numbers for both. Hypothetically a game plan for every pitcher and hitter against every hitter and pitcher. If you added in factor like location, velocity and break, the possibilities are almost endless. Hypothetically player A hits big break curves well but struggles against small break curves, etc. etc. etc. Other factors, bats R/L throws R/L. That all said, I'm looking forward to Houck's adjusted slider ranking numbers to see how well it correlates to the break ranking numbers. I'm not sure exactly what it is you're waiting for!
My to-do list on this project (more for my benefit!):
1) Redo the study with a bigger group of pitchers, including relievers.
2) It's very easy to translate actual slider results into the 20 to 80 scale (a 60 is one standard deviation better than average, a 70 is two, etc.). Translating the projected results needs a translation because the standard deviation is much lower. I have to adjust for the kurtosis (peaked-ness) of the model distribution to make it match the actual results.
3) I have an even better predictor for slider effectiveness that includes FB velocity and the percentage of pitches that are neither FB or slider (which is really low for Houck, of course). I want to look into that by substituting the average values of those two things for each pitcher, to see how his combo of FB velo and and other pitch usage contributes to effectiveness. That should be especially interesting for Houck.
4) Find a translation between R /100 pitches and wOBA allowed. Tricky, because the former includes the run value of swings and misses balls, etc. (and I believe adjusts the run value of events by the count. So if a guy misses with his curve three times and gives up a single on a 3-1 FB, much of the negative value is correctly charged to the curve).
5) Look further into my discovery of evidence for BABIP skill. It turns out that the for pitchers who throw a 4-seamer, 2-seamer, and slider, the number of guys that have bad BABIP karma (wOBA - xwOBA) on all three is exactly what you'd expect at random, but the number of guys who have good karma on all 3 pitches is decidely more (and Houck is one of them). I want to try other combinations of pitches and see if this is a general principle.
6) Armed with the above, I can make an educated adjustment for BABIP luck on Houck's slider numbers.
7) Get the benchmarks for 60, 70, 80 for all other pitch results.
8) Do a comprehensive study of FB effectiveness.
None of this will get done before the WS ends, and probably not until a lot later.
|
|
|
Post by philsbosoxfan on Oct 13, 2020 19:25:05 GMT -5
Thinking about it, it's amazing that now with the data available, you could hypothetically rank each pitchers pitch vs the quality of the competition he faced, do the same for hitters then re-run the data using adjusted numbers for both. Hypothetically a game plan for every pitcher and hitter against every hitter and pitcher. If you added in factor like location, velocity and break, the possibilities are almost endless. Hypothetically player A hits big break curves well but struggles against small break curves, etc. etc. etc. Other factors, bats R/L throws R/L. That all said, I'm looking forward to Houck's adjusted slider ranking numbers to see how well it correlates to the break ranking numbers. I'm not sure exactly what it is you're waiting for! My to-do list on this project (more for my benefit!): 1) Redo the study with a bigger group of pitchers, including relievers.
2) It's very easy to translate actual slider results into the 20 to 80 scale (a 60 is one standard deviation better than average, a 70 is two, etc.). Translating the projected results needs a translation because the standard deviation is much lower. I have to adjust for the kurtosis (peaked-ness) of the model distribution to make it match the actual results.
3) I have an even better predictor for slider effectiveness that includes FB velocity and the percentage of pitches that are neither FB or slider (which is really low for Houck, of course). I want to look into that by substituting the average values of those two things for each pitcher, to see how his combo of FB velo and and other pitch usage contributes to effectiveness. That should be especially interesting for Houck.
4) Find a translation between R /100 pitches and wOBA allowed. Tricky, because the former includes the run value of swings and misses balls, etc. (and I believe adjusts the run value of events by the count. So if a guy misses with his curve three times and gives up a single on a 3-1 FB, much of the negative value is correctly charged to the curve). 5) Look further into my discovery of evidence for BABIP skill. It turns out that the for pitchers who throw a 4-seamer, 2-seamer, and slider, the number of guys that have bad BABIP karma (wOBA - xwOBA) on all three is exactly what you'd expect at random, but the number of guys who have good karma on all 3 pitches is decidely more (and Houck is one of them). I want to try other combinations of pitches and see if this is a general principle. 6) Armed with the above, I can make an educated adjustment for BABIP luck on Houck's slider numbers. 7) Get the benchmarks for 60, 70, 80 for all other pitch results. 8) Do a comprehensive study of FB effectiveness. None of this will get done before the WS ends, and probably not until a lot later.
Getting at this part from above: I'm going to try to figure out how to take out the BABIP luck next. Whether or not I can do that, I want to get a sense of where he would rank in MLB on the three pitches.I assumed you meant in terms of results adjusted for batters faced as opposed to quality of the three pitches. You had already said his slider would rank 2nd to Klubers. Overall those two MLB wide lists SHOULD correlate closely. If they don't, there are other factors to consider. Basically what you did for Houck (the summary) but across all of baseball then rank them by pitch. LOL, not a request, just a suggestion based on how I interpreted something you typed. ADD: This would also allow you to give a numeric results value to the scouting grade for each pitch.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,931
|
Post by ericmvan on Oct 14, 2020 2:16:21 GMT -5
I'm not sure exactly what it is you're waiting for! My to-do list on this project (more for my benefit!): 1) Redo the study with a bigger group of pitchers, including relievers.
2) It's very easy to translate actual slider results into the 20 to 80 scale (a 60 is one standard deviation better than average, a 70 is two, etc.). Translating the projected results needs a translation because the standard deviation is much lower. I have to adjust for the kurtosis (peaked-ness) of the model distribution to make it match the actual results.
3) I have an even better predictor for slider effectiveness that includes FB velocity and the percentage of pitches that are neither FB or slider (which is really low for Houck, of course). I want to look into that by substituting the average values of those two things for each pitcher, to see how his combo of FB velo and and other pitch usage contributes to effectiveness. That should be especially interesting for Houck.
4) Find a translation between R /100 pitches and wOBA allowed. Tricky, because the former includes the run value of swings and misses balls, etc. (and I believe adjusts the run value of events by the count. So if a guy misses with his curve three times and gives up a single on a 3-1 FB, much of the negative value is correctly charged to the curve). 5) Look further into my discovery of evidence for BABIP skill. It turns out that the for pitchers who throw a 4-seamer, 2-seamer, and slider, the number of guys that have bad BABIP karma (wOBA - xwOBA) on all three is exactly what you'd expect at random, but the number of guys who have good karma on all 3 pitches is decidely more (and Houck is one of them). I want to try other combinations of pitches and see if this is a general principle. 6) Armed with the above, I can make an educated adjustment for BABIP luck on Houck's slider numbers. 7) Get the benchmarks for 60, 70, 80 for all other pitch results. 8) Do a comprehensive study of FB effectiveness. None of this will get done before the WS ends, and probably not until a lot later.
Getting at this part from above: I'm going to try to figure out how to take out the BABIP luck next. Whether or not I can do that, I want to get a sense of where he would rank in MLB on the three pitches.I assumed you meant in terms of results adjusted for batters faced as opposed to quality of the three pitches. You had already said his slider would rank 2nd to Klubers. Overall those two MLB wide lists SHOULD correlate closely. If they don't, there are other factors to consider. Basically what you did for Houck (the summary) but across all of baseball then rank them by pitch. LOL, not a request, just a suggestion based on how I interpreted something you typed. ADD: This would also allow you to give a numeric results value to the scouting grade for each pitch. Yeah, that's my item #6.
And it seems like BIS may have tweaked their data, because the projected effectiveness is now 4th (out of 140 starters) after Kluber, Clevinger, and Jakob Junis (who doesn't get the results to go with it). That's still top 3 percent, or about a 68 grade.
|
|
|
Post by philsbosoxfan on Oct 14, 2020 2:26:24 GMT -5
Just a quick look at Junis tells you more control factors are needed. He throws the slider 46% of the time and it's essentially the same speed as his Fastball. Easy to see why he doesn't get the same results.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,931
|
Post by ericmvan on Oct 14, 2020 11:23:18 GMT -5
Just a quick look at Junis tells you more control factors are needed. He throws the slider 46% of the time and it's essentially the same speed as his Fastball. Easy to see why he doesn't get the same results. I have 91.6 FB, 81.9 SL for Junis. What year are you looking at?
This conversation has inspired a great idea to pursue if I ever go deeper into these models. The non-random difference between the results and the prediction has to be command of the pitch, right? But you can't really put BB% or Strike% into the model because that's part of the results. (And you can't use % of sliders as a variable because that's a function of how good you think your slider is, which correlates to how good it actually is.)
What you can do, however, is look at the set of differences, pitch by pitch, of each pitcher. Do they tend to correlate with each other? The average differential should be a measure of overeall command.
And then I can look at how that relates to overall control. I already know from a study I did nine years ago that there's a tremendously strong relationship between BB%, Strike%, and P/PA, which makes perfect sense, and a very small but highly significant relationship of P/PA to BABIP, which is lower for guys who throw more P/PA (which also makes sense. In the same study I discovered that guys who threw more changeups / splitters had a fractionally lower BABIP).
I really have to get back up to speed with ANOVA (beginning with taking a look at what's included in the current version of Excel, which I have on a new laptop while I'm running Excel 2010 on this one). With that, I can use each season of each pitcher and use the within-pitcher differences as well as the between-pitcher differences. So "redo every model with ANOVA" should be on the to-do list!"
|
|
|
Post by philsbosoxfan on Oct 14, 2020 19:49:46 GMT -5
My bad, wrong column on the velocities for Junis.
I'm surprised that frequency of use wouldn't be a factor and a large one at that. There has to be a batter expectation component. Taking it to the extreme, if Junis was a one pitch pitcher, I'm pretty sure the results would decline significantly. I was actually psychologically keying more on the 46%. If you were to guess slider, you'd be right almost 50% of the time. Astros territory.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,931
|
Post by ericmvan on Oct 15, 2020 2:47:54 GMT -5
My bad, wrong column on the velocities for Junis. I'm surprised that frequency of use wouldn't be a factor and a large one at that. There has to be a batter expectation component. Taking it to the extreme, if Junis was a one pitch pitcher, I'm pretty sure the results would decline significantly. I was actually psychologically keying more on the 46%. If you were to guess slider, you'd be right almost 50% of the time. Astros territory. Frequency of use is presumably a large component in a predictive regression. The problem is, is it a cause or an effect? Is your slider better because you throw it more often (probably, yes, because you command it better), or do you throw it more often that someone with a slider that isn'ta s good (almost certainly yes).
I did include it in a changeup analysis that I did back in 2008, and it was really big, just as you'd expect if it were both cause and effect. I spent a ton of time trying to figure out how I might separate the two, without coming up with an idea worth trying.
I will try adding it to the current model to see what happens -- getting the size of the contribution for each pitch will be interesting. FB usage has been decreasing the last few years and its effectiveness has been going up, so you are correct in that there's game theory involved and an optimum set of frequencies for your pitches, depending on how effective each one is.
|
|
|
Post by philsbosoxfan on Oct 15, 2020 3:02:56 GMT -5
Not to keep you up at night but I was also thinking about break consistency. If all a pitchers sliders are essentially the same, it would seem to be a disadvantage. Using Miriano as an example, everyone pretty much knew a cutter was coming, the question was which cutter was coming. That would seem to be a game by game thing, can pitchers control break during a game, during a single at bat ? It seems your break data is an average, not a range.
With tunneling being the new thang on the block and release point data being available for each pitch, I would also assume that could also be factored in. I know that Abbott mentioned it earlier in the summer when he was talking about Seabold so, I know it's out there more than just the Pitching Ninja.
Haha, go back to sleep.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,931
|
Post by ericmvan on Oct 15, 2020 12:09:59 GMT -5
Not to keep you up at night but I was also thinking about break consistency. If all a pitchers sliders are essentially the same, it would seem to be a disadvantage. Using Miriano as an example, everyone pretty much knew a cutter was coming, the question was which cutter was coming. That would seem to be a game by game thing, can pitchers control break during a game, during a single at bat ? It seems your break data is an average, not a range. With tunneling being the new thang on the block and release point data being available for each pitch, I would also assume that could also be factored in. I know that Abbott mentioned it earlier in the summer when he was talking about Seabold so, I know it's out there more than just the Pitching Ninja. Haha, go back to sleep. Somewhere here I already concluded that "changing movement" should be as useful as changing speeds. But that would be a bitch to measure. You have to determine the point at where less movement on a slider or curve becomes "hanging" the pitch and backfires, and then calculate the variance of good movement only. I'm not going there unless someone pays me!
But as this is a skill, not a tool, for our purposes we can ignore it. But it's also another way you can outperform or underperform your projection based on average velo and movement. Change speeds, change movement, and command it, and you'll be much better than expected. Do none of those things, and you'll be much worse.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,931
|
Post by ericmvan on Oct 23, 2020 15:05:34 GMT -5
(First of two parts, because I forgot to hit "Create Post" last night!)
So I just realized that Houck's pitch values are measured in runs per 100 pitches! That makes an adjustment for the quality of his opposition trivial.
So, adjusting every pitch he threw to every hitter, based on how well that hitter hit that pitch all year ... the average pitch he threw was to a hitter produced 0.66 runs above average per 100 pitches.
He threw 257 pitches, so that's 1.7 runs above average in 17 innings of work, which is 1.6 ER, which is 0.82 points of ERA.
So his opponent-adjusted xERA, and his pERA (my metric) are 3.01. That's a .259 xwOBA. And that does not credit him for his skill in inducing GDP's (the big weakness in wOBA and xwOBA for pitchers).
How good is that? Adjusting for virtual league:
2.52 Bauer
2.72 DeGrom 2.93 Lamet
2.97 Fried
3.01 Hauck
3.06 Bieber 3.07 Javier
3.09 Means 3.09 Sanchez, Sixto
3.11 Cole
3.12 Bundy
|
|
gerry
Veteran
Enter your message here...
Posts: 1,667
|
Post by gerry on Oct 23, 2020 17:53:48 GMT -5
So he is as good as he appeared to be. Thank you.
BTW, pera is the Spanish pronunciation for an amazing fruit: the pear. Jacoby Ellsbury’s ancestors thrived on peras/pears and peaches until conquering zero Kit Carson and the cavalry boxed the Navajo resistance in Canyon de Shelly (Shay) and burnt down all the fruit trees, leading to their surrender and subsequent long walk. Peras are resilient and important.
|
|
|
Post by iakovos11 on Oct 23, 2020 19:05:17 GMT -5
(First of two parts, because I forgot to hit "Create Post" last night!)
So I just realized that Houck's pitch values are measured in runs per 100 pitches! That makes an adjustment for the quality of his opposition trivial. So, adjusting every pitch he threw to every hitter, based on how well that hitter hit that pitch all year ... the average pitch he threw was to a hitter produced 0.66 runs above average per 100 pitches. He threw 257 pitches, so that's 1.7 runs above average in 17 innings of work, which is 1.6 ER, which is 0.82 points of ERA.
So his opponent-adjusted xERA, and his pERA (my metric) are 3.01. That's a .259 xwOBA. And that does not credit him for his skill in inducing GDP's (the big weakness in wOBA and xwOBA for pitchers).
How good is that? Adjusting for virtual league:
2.52 Bauer
2.72 DeGrom 2.93 Lamet
2.97 Fried
3.01 Hauck
3.06 Bieber 3.07 Javier
3.09 Means 3.09 Sanchez, Sixto
3.11 Cole
3.12 Bundy
So we're still talking about the incredibly insane Small Sample Size of his MLB innings in 2020, right? Just massaging those numbers in a different way.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,931
|
Post by ericmvan on Oct 24, 2020 2:49:37 GMT -5
(First of two parts, because I forgot to hit "Create Post" last night!)
So I just realized that Houck's pitch values are measured in runs per 100 pitches! That makes an adjustment for the quality of his opposition trivial. So, adjusting every pitch he threw to every hitter, based on how well that hitter hit that pitch all year ... the average pitch he threw was to a hitter produced 0.66 runs above average per 100 pitches. He threw 257 pitches, so that's 1.7 runs above average in 17 innings of work, which is 1.6 ER, which is 0.82 points of ERA.
So his opponent-adjusted xERA, and his pERA (my metric) are 3.01. That's a .259 xwOBA. And that does not credit him for his skill in inducing GDP's (the big weakness in wOBA and xwOBA for pitchers).
How good is that? Adjusting for virtual league:
2.52 Bauer
2.72 DeGrom 2.93 Lamet
2.97 Fried
3.01 Hauck
3.06 Bieber 3.07 Javier
3.09 Means 3.09 Sanchez, Sixto
3.11 Cole
3.12 Bundy
So we're still talking about the incredibly insane Small Sample Size of his MLB innings in 2020, right? Just massaging those numbers in a different way. Just trying to pin down how well he actually pitched. In a small sample, it's possible to face a lineup that was way better than average. There's a huge difference between a 3.01 true ERA and a 3.83.
In terms of the meaningfulness of being that good over your first 17 MLB innings ... pitch movement and velocity stabilize really, really quickly, and in Houk's case one of his three starts appears to have been a significant off-night, so that inevitable regression is probably already factored in. And what we know is that he seems to have the velocity and movement on both his slider and sinker to project as 70 pitches (the sinker analysis should be up tomorrow), and of course he fared way better than that with them in these starts.
Did anyone on the staff think that Houck could display a 70 slider and a 70 sinker over his first three MLB starts? It's a plain fact that he did, and I don't believe he was supposed to be able to do that ever, let alone immediately.
|
|
|
Post by manfred on Oct 24, 2020 11:45:27 GMT -5
So we're still talking about the incredibly insane Small Sample Size of his MLB innings in 2020, right? Just massaging those numbers in a different way. Just trying to pin down how well he actually pitched. In a small sample, it's possible to face a lineup that was way better than average. There's a huge difference between a 3.01 true ERA and a 3.83.
In terms of the meaningfulness of being that good over your first 17 MLB innings ... pitch movement and velocity stabilize really, really quickly, and in Houk's case one of his three starts appears to have been a significant off-night, so that inevitable regression is probably already factored in. And what we know is that he seems to have the velocity and movement on both his slider and sinker to project as 70 pitches (the sinker analysis should be up tomorrow), and of course he fared way better than that with them in these starts.
Did anyone on the staff think that Houck could display a 70 slider and a 70 sinker over his first three MLB starts? It's a plain fact that he did, and I don't believe he was supposed to be able to do that ever, let alone immediately.
I like all the positives you are raising, but just to tap the breaks: they didn’t expect it because in his longer record, he hadn’t done it. So the question becomes are those few games now the truest indicator of who he is? Are they peak performance that might show itself often? Sometimes? Rarely? Obviously I don’t have an answer. It is obviously awesome he showed a plus slider especially... if he did it then, he can do it again. But the question of sustaining it remains. I remain optimistic, don’t get me wrong. He is far more exciting than most of what I saw in 2020.
|
|
|
Post by philsbosoxfan on Oct 24, 2020 12:17:55 GMT -5
Just trying to pin down how well he actually pitched. In a small sample, it's possible to face a lineup that was way better than average. There's a huge difference between a 3.01 true ERA and a 3.83. In terms of the meaningfulness of being that good over your first 17 MLB innings ... pitch movement and velocity stabilize really, really quickly, and in Houk's case one of his three starts appears to have been a significant off-night, so that inevitable regression is probably already factored in. And what we know is that he seems to have the velocity and movement on both his slider and sinker to project as 70 pitches (the sinker analysis should be up tomorrow), and of course he fared way better than that with them in these starts. Did anyone on the staff think that Houck could display a 70 slider and a 70 sinker over his first three MLB starts? It's a plain fact that he did, and I don't believe he was supposed to be able to do that ever, let alone immediately.
I like all the positives you are raising, but just to tap the breaks: they didn’t expect it because in his longer record, he hadn’t done it. So the question becomes are those few games now the truest indicator of who he is? Are they peak performance that might show itself often? Sometimes? Rarely? Obviously I don’t have an answer. It is obviously awesome he showed a plus slider especially... if he did it then, he can do it again. But the question of sustaining it remains. I remain optimistic, don’t get me wrong. He is far more exciting than most of what I saw in 2020. I think it was Dykstra that pointed out that all six of his hits were against the fastball. Don't count out the sinker. Also don't forget Abbott said he has a cutter but just didn't need it so he didn't use it. I know in this case it's both ends of the equation but would you have more confidence in a SSS that was excellent results with so-so stuff or a pitcher that had so-so results with excellent stuff ?
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,931
|
Post by ericmvan on Oct 24, 2020 12:29:43 GMT -5
Just trying to pin down how well he actually pitched. In a small sample, it's possible to face a lineup that was way better than average. There's a huge difference between a 3.01 true ERA and a 3.83.
In terms of the meaningfulness of being that good over your first 17 MLB innings ... pitch movement and velocity stabilize really, really quickly, and in Houk's case one of his three starts appears to have been a significant off-night, so that inevitable regression is probably already factored in. And what we know is that he seems to have the velocity and movement on both his slider and sinker to project as 70 pitches (the sinker analysis should be up tomorrow), and of course he fared way better than that with them in these starts.
Did anyone on the staff think that Houck could display a 70 slider and a 70 sinker over his first three MLB starts? It's a plain fact that he did, and I don't believe he was supposed to be able to do that ever, let alone immediately.
I like all the positives you are raising, but just to tap the breaks: they didn’t expect it because in his longer record, he hadn’t done it. So the question becomes are those few games now the truest indicator of who he is? Are they peak performance that might show itself often? Sometimes? Rarely? Obviously I don’t have an answer. It is obviously awesome he showed a plus slider especially... if he did it then, he can do it again. But the question of sustaining it remains. I remain optimistic, don’t get me wrong. He is far more exciting than most of what I saw in 2020. When guys who have great stuff fall short of expectations, it's very rare that it's because they lose their stuff, without that being a consequence of health. I can't think of an example, in fact -- probably because if a guy's stuff starts declining, we assume it's health-related!
What we don't know is where Houck's command of his stuff ranked in these three starts relative to the average command he will have going forward. When guys with great stuff fail to realize their upside even though they stay healthy, it's almost always because they can't command it well enough.
My argument here is that the odds of him reaching his 6 upside are much higher than his current ranking indicates. (He should also go from a 4.5 projection to a 5).
|
|
|
Post by manfred on Oct 24, 2020 12:31:31 GMT -5
I like all the positives you are raising, but just to tap the breaks: they didn’t expect it because in his longer record, he hadn’t done it. So the question becomes are those few games now the truest indicator of who he is? Are they peak performance that might show itself often? Sometimes? Rarely? Obviously I don’t have an answer. It is obviously awesome he showed a plus slider especially... if he did it then, he can do it again. But the question of sustaining it remains. I remain optimistic, don’t get me wrong. He is far more exciting than most of what I saw in 2020. I think it was Dykstra that pointed out that all six of his hits were against the fastball. Don't count out the sinker. Also don't forget Abbott said he has a cutter but just didn't need it so he didn't use it. I know in this case it's both ends of the equation but would you have more confidence in a SSS that was excellent results with so-so stuff or a pitcher that had so-so results with excellent stuff ? I am optimistic, as I said. But I also look on those results as probably an aberration — but they might point to a maximum potential he can hit.
|
|
|
Post by philsbosoxfan on Oct 24, 2020 12:46:15 GMT -5
There's obviously no way he's going to maintain a .053 ERA. #3 starter would be far more than was expected before he made it to Boston and he'd have to regress significantly to get to #3. To me #1,#2 is my optimistic view, #3 my present view and #4 or #5 my pessimistic view. If nothing else, he'll be interesting to see how it shakes out. The only negative is that he'll likely only get 150 innings max but nowadays, that's not all that bad.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,931
|
Post by ericmvan on Oct 24, 2020 12:55:11 GMT -5
So his opponent-adjusted xERA, and his pERA (my metric) are 3.01. That's a .259 xwOBA. A nd that does not credit him for his skill in inducing GDP's (the big weakness in wOBA and xwOBA for pitchers).He faced 13 guys in GDP situations.
He struck out 6 and walked 1.
Of the other 6 ... four hit into DP's.
Sneak preview: my sinker model has him as an 85, not a 70 (yes, in theory the 20-80 scale can go past 80).
The cost is a 40 movement / velo on his 4-seamer; part 3 of this will be an examination of how he got such good results with it.
|
|
|
Post by manfred on Oct 24, 2020 13:31:33 GMT -5
So his opponent-adjusted xERA, and his pERA (my metric) are 3.01. That's a .259 xwOBA. A nd that does not credit him for his skill in inducing GDP's (the big weakness in wOBA and xwOBA for pitchers).He faced 13 guys in GDP situations.
He struck out 6 and walked 1.
Of the other 6 ... four hit into DP's.
Sneak preview: my sinker model has him as an 85, not a 70 (yes, in theory the 20-80 scale can go past 80).
The cost is a 40 movement / velo on his 4-seamer; part 3 of this will be an examination of how he got such good results with it.
I’m not trying to be difficult, cause I like Houck and hope he’s great, but I wonder if similar pitchers (say Masterson or worse) have strung together similar stretches that flashed stuff beyond what they otherwise sustain? In other words, is this stretch, short as it is, so rare an uptick that it becomes statistically meaningful?
|
|
|
Post by umassgrad2005 on Oct 24, 2020 14:12:39 GMT -5
Just my two cents, Houcks issues were never a lack of arm talent. When we drafted him guys said he had 1.5 pitches because his slider would come and go. He then added a splitter this year, which was great. It's not that he flashed stuff we didn't know he had, it's that he maintained it threw games. Can he do that without months of sim games? What will hitters do because they now have tape on him?
Hitters are going to get more hits against him, that's just a given. His walks were still an issue having 3 in each start. Hitter aren't going to come close to hitting .113 .443 OPS against him, he won't have a .161 BAbip, like not even close. In a ton of ways what Houck just did was what Chavis did when he first got called up. Now let's hope he builds on this, keeps getting better. Yet he's a #5 for me next year till he proves himself and I don't think he just did that, more like he earned a rotation spot.
|
|
|