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.
9/10-9/13 Red Sox @ Rays Series Thread
|
Post by Chris Hatfield on Sept 14, 2020 10:39:17 GMT -5
You guys know by now that I love to calculate the odds of splits being random. So, before 2019 Christian Arroyo had 26 professional home runs in 2006 career PA. He now has 12 in his last 209. The odds of getting that split in a random simulation based on his total career HR rate are ... wait for it .... 400,178 to 1. Continued later today in the 2021 thread, where he (and Munoz) have quickly become complicating factors of the kind you like to have.
Eric, since you're leaning pretty hard into this Arroyo thing, are you controlling at all for the fact that AAA last year was using a different baseball? There was a HR spike across AAA last year that surely accounts for some of Arroyo's power. Looking further, aggregating his entire previous 2006 PA ignores that he's hit for some power when he's been in hitter-friendly leagues. He had 9 in his season in the California League in 381 AB. He had 4 in just 91 AB in the PCL in 2017. I guess I'd like to see a bit more digging and controlling for the league before I get excited about his power spike last year. And if this is in some other post, I apologize, but I know I've seen you posting about Arroyo's 2019 HR totals elsewhere as well and didn't recall any controlling for the ball.
|
|
|
Post by umassgrad2005 on Sept 14, 2020 11:53:07 GMT -5
Thank you, Baseball players aren't just numbers and if you want to try and determine something isn't random you can't treat the stats like they are just numbers without context. Otherwise every hot streak looks like a miracle. It's not rare for non HR hitters to have hot streaks and get you excited, Peraza say hello.
Arroyo has shown much more power in the minors than a guy like Peraza. It's not uncommon for a guy like him to turn some doubles into HRs as he gets older in the year of the HR. Put another way I don't think the odds are that bad that a guy at age 20 who hit 9 HR in 409 PA would have a stretch like he has at age 24/25.
I actually like Arroyo, not because of his homer streak, put because for years he actually showed you much more power than a guy like Peraza. No matter what he does that was a nice pickup by Bloom.
|
|
gerry
Veteran
Enter your message here...
Posts: 1,660
|
Post by gerry on Sept 14, 2020 13:48:08 GMT -5
Christian Arroyo. I don’t have the stat skills to make a good case for anyone. I do know that the primary weakness of this 2020 (and 2019) team has been pitching, and will be again In 2021 if not aggressively addressed.
I also know that 2B for ~2022 forward is likely J. Downs. Therefore, we don’t want or need to spend either $$ or prospects on a 2B like B. Lowe, though that would, any other time, be a good thing. We need pitching.
IMO virtually all of the in house, low cost 2B candidates for 2021 (Arroyo, Chatham, Chavis, Lin, Muñoz, Peraza) would be acceptable even on a contending team. But Arroyo, with consistent playing time, seems to have the potential to live up to the expectations which made him a first round pick in the first place. He seems to be the leading candidate right now. Am I correct?
|
|
redsox04071318champs
Veteran
Always hoping to make my handle even longer...
Posts: 15,644
Member is Online
|
Post by redsox04071318champs on Sept 14, 2020 14:16:25 GMT -5
Thank you, Baseball players aren't just numbers and if you want to try and determine something isn't random you can't treat the stats like they are just numbers without context. Otherwise every hot streak looks like a miracle. It's not rare for non HR hitters to have hot streaks and get you excited, Peraza say hello. Arroyo has shown much more power in the minors than a guy like Peraza. It's not uncommon for a guy like him to turn some doubles into HRs as he gets older in the year of the HR. Put another way I don't think the odds are that bad that a guy at age 20 who hit 9 HR in 409 PA would have a stretch like he has at age 24/25. I actually like Arroyo, not because of his homer streak, put because for years he actually showed you much more power than a guy like Peraza. No matter what he does that was a nice pickup by Bloom. Exactly - do you remember when Brock Holt suddenly started hitting home runs in Sept 2018? It's like he had 1 and then whacked 6 just like that? The following year, he was back to his usual powerless (but fun to watch) self. Streaky inconsistencies happen. Hell, I remember Mike Benjamin piling up multiple hit games for what felt like a week in 1998, but the guy was still a lousy hitter, the same afterward as he was before then. Stuff happens. Most of it is flukey. Once in a while you get a Jose Bautista story, but most of the time not.
|
|
redsox04071318champs
Veteran
Always hoping to make my handle even longer...
Posts: 15,644
Member is Online
|
Post by redsox04071318champs on Sept 14, 2020 14:21:30 GMT -5
Christian Arroyo. I don’t have the stat skills to make a good case for anyone. I do know that the primary weakness of this 2020 (and 2019) team has been pitching, and will be again In 2021 if not aggressively addressed. I also know that 2B for ~2022 forward is likely J. Downs. Therefore, we don’t want or need to spend either $$ or prospects on a 2B like B. Lowe, though that would, any other time, be a good thing. We need pitching. IMO virtually all of the in house, low cost 2B candidates for 2021 (Arroyo, Chatham, Chavis, Lin, Muñoz, Peraza) would be acceptable even on a contending team. But Arroyo, with consistent playing time, seems to have the potential to live up to the expectations which made him a first round pick in the first place. He seems to be the leading candidate right now. Am I correct? I would say not really. All of those guys you mention are at best utility players (Munoz could be a decent useful utility guy), not guys that contending teams would want to start at 2b. None of them can carry a .360 OBP the way Holt did. Maybe, and I say, maybe Arroyo could be a different story - he has a non-zero chance of being a legit regular. I was hoping Chavis could be the 2b, but his inability to catch up to plus velocity is a legit problem. The Red Sox can do better at 2b than the guys you listed. I suspect they'll get a hold-the-fort guy until Downs is ready. There is a non-zero chance it's Arroyo.
|
|
|
Post by Chris Hatfield on Sept 14, 2020 14:32:45 GMT -5
Thank you, Baseball players aren't just numbers and if you want to try and determine something isn't random you can't treat the stats like they are just numbers without context. Otherwise every hot streak looks like a miracle. It's not rare for non HR hitters to have hot streaks and get you excited, Peraza say hello. Arroyo has shown much more power in the minors than a guy like Peraza. It's not uncommon for a guy like him to turn some doubles into HRs as he gets older in the year of the HR. Put another way I don't think the odds are that bad that a guy at age 20 who hit 9 HR in 409 PA would have a stretch like he has at age 24/25. I actually like Arroyo, not because of his homer streak, put because for years he actually showed you much more power than a guy like Peraza. No matter what he does that was a nice pickup by Bloom. To push back in the other direction a bit, that's what Eric is talking about when he says he's looking at whether splits are random. I wasn't trying to make a point to push back. I was trying to point out another potential factor that needed to be controlled for an analysis to be robust enough, imo, to tell us anything. Not every post on this board is in full agreement or full disagreement with the post it's quoting. It's a discussion forum not a debate forum.
|
|
|
Post by umassgrad2005 on Sept 14, 2020 14:36:26 GMT -5
Bloom needs to make a plan and give guys a long look. Like don't treat young guys like they did Chavis before the Moreland trade. It's not surprising he's now doing better with actual regular playing time. If Bloom believes in Arroyo then give him a long run to see what you have.
Yet if I had to guess he brings in a cheap veteran and let's them battle it out in spring training.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,923
|
Post by ericmvan on Sept 15, 2020 2:09:06 GMT -5
You guys know by now that I love to calculate the odds of splits being random. So, before 2019 Christian Arroyo had 26 professional home runs in 2006 career PA. He now has 12 in his last 209. The odds of getting that split in a random simulation based on his total career HR rate are ... wait for it .... 400,178 to 1. Continued later today in the 2021 thread, where he (and Munoz) have quickly become complicating factors of the kind you like to have.
Eric, since you're leaning pretty hard into this Arroyo thing, are you controlling at all for the fact that AAA last year was using a different baseball? There was a HR spike across AAA last year that surely accounts for some of Arroyo's power. Looking further, aggregating his entire previous 2006 PA ignores that he's hit for some power when he's been in hitter-friendly leagues. He had 9 in his season in the California League in 381 AB. He had 4 in just 91 AB in the PCL in 2017. I guess I'd like to see a bit more digging and controlling for the league before I get excited about his power spike last year. And if this is in some other post, I apologize, but I know I've seen you posting about Arroyo's 2019 HR totals elsewhere as well and didn't recall any controlling for the ball. I haven't controlled for league / year or park, but there has also been a statistically significant increase in K% + BB% (slugger profile) starting last year. And, yes, the new baseball would help him ... but not help him hit homers five times as often. By now you've probably seen the 900,000+ to 1 odds against that season being random. And he's sustained the Durham home run rate in MLB.
Basically, this is one of the rare occasions where the effect size is so large that it's safe to assume that the hitter is responsible for a decent chunk of it. Nobody thinks he's a 35 HR guy but he's likely 10-20, where his best prospect projection was 15.
I think that some people are missing this ... when a guy has a huge effect size in a small sample and it ends up as a ridiculously significant split, the huge odds against it being random simply give you confidence that it's real in some fashion. It's not an argument that he's anywhere as good as his numbers. The classic example is from Bill James: Clemens fanning 15 and walking 0 as a rookie. That didn't mean that he was going to sustain numbers anywhere near that. But it established a very high floor.
Umassgrad actually has the right take, except that he's mistaking the good HR years in the minors as meaningful when they were probably just the park! What these numbers do is raise Arroyo's HR / year floor and ceiling from bad to decent (the ceiling maybe sneaking into the good range), but more importantly, it's essentially impossible to explain the numbers unless this has really happened. It's the latter that's important.
Another thing to keep in mind: as UMG points out, MLB history is full of guys who had crazy little homer streaks and never hit like that again. Sometimes a guy changes their approach and turns his cold zones into hot ones (while creating new cold zones to exploit) and goes off on a tear until teams get a handle on it. I was at the game where Kevin Millar hit 3 homers against the Yankees with his new radically open stance, the day before the legendary ARod / 'Tek / Mo / Bill Mueller game. That year Millar had 5 HR in 244 contacts, then 6 in 19 after changing his stance; the odds against that are 1.7 billion to 1! But it was real. After you eliminated all the cases like this, you'd be left with a few clusters that were truly random, exactly as many as you'd expect given how much baseball had been played. But an alteration over two years for three different teams, when compared to an entire previous career -- that's not the same thing at all.
Let's see ... He did manage an .029 HR/C in San Jose and .051 (102 PA) in Sacramento in the PCL. But his walk rate was .044 and .050; it was .090 in Durham. You know that BB rate and power are strongly correlated (with causal arrows going both ways). If you look at his HRC for every partial year, even with the wide variance caused by the different ml parks and the small samples in MLB in Tampa and Boston, you still get a statistically different difference.
OK ... the HRC in the IL went from .032 to .050 from 2018 to 2019, a startling 57% increase. Arroyo played both years in Durham. He had a .014 his first year there. He wouldn't be expected to benefit more than the average guy (in fact, I think the guys who benefit most from a juiced ball are guys with average power), so you're looking for him to put up an .023 rate. He actually had an .084. He had hit 2 homers in 138 times making contact and the juiced ball was going to let him him 2 again on just 95 contacts. But he hit 8. That happens 10% of the time at random if you ignore his entire career before those two years, all of which make you hugely more certain that the 2 HR was not a fluke.
So let's just ball-adjust the 8 HR at Durham to 5, while leaving the park-inflated Cal and PCL figures in the comparison. (So the homer spike was half him and half the ball, and I almost wrote "responsible for maybe half of it" before deciding that "a decent chunk" was better.) It's still 1 in 35,875 against it being random, factoring in the rise in K rate and BB rate (taken together; if I did them independently it would be much higher).
BTW, I wouldn't yet pencil him in for anything next year. But he's earning a spot on the 40-man.
|
|
|
Post by philsbosoxfan on Sept 15, 2020 2:44:25 GMT -5
No options so you have to pencil him into something.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,923
|
Post by ericmvan on Sept 15, 2020 3:26:10 GMT -5
No options so you have to pencil him into something. You might trade him at the end of ST if he proves to not make roster sense. There's a scenario where you have him and Chavis as the backup MI candidates and Chavis looks like he's made a step forward of some sort and you simply can't option him, and someone is willing to give you a nice lottery ticket for Arroyo.
|
|
|
Post by philsbosoxfan on Sept 15, 2020 3:29:48 GMT -5
No options so you have to pencil him into something. You might trade him at the end of ST if he proves to not make roster sense. There's a scenario where you have him and Chavis as the backup MI candidates and Chavis looks like he's made a step forward of some sort and you simply can't option him, and someone is willing to give you a nice lottery ticket for Arroyo. True but options will be part of the equation. Lin also has no options but he's unlikely to make the Rule 'V' cut.
|
|
|
Post by philsbosoxfan on Sept 15, 2020 3:38:55 GMT -5
The other wildcard here is Munoz who does have an option. If he doesn't beat out Benni for the starting LF job, he will be in the mix for either the 2B or utility role.
|
|
|
Post by umassgrad2005 on Sept 15, 2020 12:44:27 GMT -5
A few things, Baseball numbers for players are random numbers, they go up, they go down, they are all over the place. Trying to calculate the odds of random numbers being random needs a ton of adjustment and context no?
They aren't non random numbers like the amount of milk a family buy weekly or how many hours a person watches TV daily. Huge spikes in things that aren't random have meaning. The average US household income, the debt amount per person, asset to debt ratio for businesses. The price of GE stock and it's huge drop in price, that's crazy meaningful because stock prices aren't just random numbers right? My Banks big one is earnings per share, like literally the only thing they care about.
I love the Bill James piece because he's looking at it like I do. It's not trying to calculate odds that game wasn't random. He's adding context, as in not many players in Baseball history have every had 15 strikeout games with zero walks. That's what is giving it meaning. That's what is raising a players floor, the context of comparing that to the history of Baseball.
So if you want to look at guys like Holt, Peraza, and Arroyo to see if small bursts of power are meaningful don't you need to look at Baseball History? How many player per year have done things like that over the last 100 years or focus on maybe smaller samples like the last 20 years? Maybe since MLB started testing? Heck even then you likely need to adjust a bunch of things, Coors Field and crap like that.
|
|
|
Post by philsbosoxfan on Sept 15, 2020 12:50:25 GMT -5
They are only random numbers in Strat-O-Matic baseball. If they were random, the odds on Mookie having a good year would be the same as Chris Davis.
|
|
|
Post by umassgrad2005 on Sept 15, 2020 13:34:18 GMT -5
They are only random numbers in Strat-O-Matic baseball. If they were random, the odds on Mookie having a good year would be the same as Chris Davis. Betts is likely the second most consistent player in Baseball and his offensive bwar totals have been 4.8, 6.4, 3.3, 8.7 and 5.2. You start breaking down small sample sizes of his numbers and he'll have all types of streaks from horrible to great. The numbers are random even with Betts, just in a different way from other players. We are talking small sample sizes right? No Baseball player plays the same way every game, week, month or year right? The difference in Betts vs Davis isn't small samples, it large sample sizes, because I'm sure even Chris Davis has had same sample sizes were he looks like a great hitter no?
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,923
|
Post by ericmvan on Sept 15, 2020 15:02:07 GMT -5
They are only random numbers in Strat-O-Matic baseball. If they were random, the odds on Mookie having a good year would be the same as Chris Davis. Betts is likely the second most consistent player in Baseball and his offensive bwar totals have been 4.8, 6.4, 3.3, 8.7 and 5.2. You start breaking down small sample sizes of his numbers and he'll have all types of streaks from horrible to great. The numbers are random even with Betts, just in a different way from other players. We are talking small sample sizes right? No Baseball player plays the same way every game, week, month or year right? The difference in Betts vs Davis isn't small samples, it large sample sizes, because I'm sure even Chris Davis has had same sample sizes were he looks like a great hitter no? You're overthinking things here. It's really simple.
A player shows a change in performance over a certain sample size.
To begin with, you want to know, to what degree is this worth our attention?
Calculating the odds that this would happen in a a random simulation answers that question beautifully.
The next question is the one that really matters: is this change predictive for the future?
The random odds do not address that question the tiniest bit. All they do is tell you how much time you should spend investigating or considering the possibility that it's real.
So when I say, wow, the odds of Joe X hitting this much better over the last Time Interval are Really Big Number to 1, I'm not saying anything about whether Joe X has actually become better. All I'm saying is:
- This is worth looking into if you're a number cruncher or scouting maven. What do those perspectives tell us?
- This guy is worth including in your speculations for the future if you're playing GM. You should not leave Joe X unmentioned; you should say things like "if Joe X is for real, then we should instead do this ..."
I believe that in most cases where the odds are long, something real is happening. Far and away the biggest reason for the change not being sustainable is that the opposition adjusts. And it's really hard to guess how much that will cut into the gains.
Now, in Arroyo's case I found a good reason to think he was truly better, because his K and BB rates had changed at the same time, in just the way you'd expect of he had turned himself into more of a power hitter. So he's not a good example for this explanation.
So let's move on to Munoz. I haven't crunched any such numbers for him at all! And I'm trying very hard not to do so.
But if, in a moment of weakness I do, all the odds numbers will tell you is whether we should list him with the starting 2B candidates or not.
I will always let you know when I think odd numbers represent something real. All the odd numbers alone mean is "attention should be paid."
|
|
|
Post by manfred on Sept 15, 2020 15:18:00 GMT -5
They are only random numbers in Strat-O-Matic baseball. If they were random, the odds on Mookie having a good year would be the same as Chris Davis. Easy on strat-o-matic! When I was a kid, I’d test players by doing their full season of at bats or starts, and the outcomes were remarkably accurate! I always thought that was so cool (there were exceptions... always good to have invincibles like Don Carmen).
|
|
|