• Content count

  • Joined

  • Last visited

  • Days Won


tonycpsu last won the day on August 6

tonycpsu had the most liked content!

Community Reputation

1,056 Excellent

About tonycpsu

  • Rank

Previous Fields

  • Add to Mailing List?
  1. [A couple comments removed. Disagree without being disagreeable, guys. That means, at a minimum, no age-based attacks on other members.]
  2. I also prioritized Chatwood. He's not a lock, but of widely-available streamers, he's about as good as it gets this late in the season. Also grabbed Marquez. My backups were Junis and Tomlin, and I'm glad I don't have to think about rolling with those guys.
  3. I watch for the human element of athletes doing amazing things on the field, not the human element of doughy middle-aged dudes like me trying to accomplish a task that human eyesight and reaction time is ill-suited for. But, hey, you do you!
  4. My problem is that you're using categorical terms like "everyone" and "infallible" that clearly don't apply to this audience. The point about the "Fox Trax" box or whatever might be a compelling argument in an audience of casual baseball fans, but this is a forum dedicated to people who obsess over details of the game that others barely notice. It's September 19th, and many leagues are already decided, yet there's a handful of people who care about the issue of getting ball/strike calls right that they're participating in this discussion. Do you really think people here aren't aware that the chart on TV is an approximation? Perhaps there are some, but I think it'd be much more sporting of you to engage with the prevailing opinion when you're making blanket statements about "everyone" feeling a certain way, especially when I've gone out of my way to acknowledge the limitations of the technology. The ultimate standard is the MLB rulebook. The tech is designed around those rules, and MLB and the umps both agree that it's the standard. Absent a conspiracy that involves the MLB, the umpires, and/or possibly some secret hacking group, I have confidence that the system in place is measuring strikes according to the rulebook definition more accurately than human umpires are. If you disagree, then you're on the opposite side of the league and the umpires themselves, and I guess we'll agree to disagree.
  5. This is my final request for you to cite an example of someone in this thread actually expressing this naive attitude before I conclude that you just want to remain on your soapbox instead of engaging in a discussion. You're more than welcome to continue repeating these same opinions over and over, but you keep citing a "computers must be perfect" attitude that, as far as I can tell, nobody has expressed, and there's a point at which I have to conclude it's a willful misrepresentation instead of a failure to understand. From the FiveThirtyEight piece I linked to above: The chart shows that even Barksdale's accuracy in 2015 topped out at 90%, which means one in ten of his ball/strike calls was wrong as compared to the rulebook zone captured by the computers. Strictly speaking, with ~300 pitches per game, that means they're really missing 30 calls a game, or about 3 per inning. Of course that just means they're "incorrect" compared to what the computer says the zone is. The computer can be wrong, as I've acknowledged earlier. But MLB is using the tech to push umpires in the direction of calling the zone that way. The Brooks Baseball charts try to account for this by superimposing the "typical" zone of an MLB ump on top of the rulebook zone -- the umps' typical zones are a little wider on both sides for right-handed hitters, and a lot wider on just the outside edge for let-handed hitters. And, hey, maybe the umps are right that this is the kind of zone that should be called, with more room on the corners! If that's the case, the computers can adjust for that. But, based on what the system says a strike is right now, they're missing well more than a call per inning on average.
  6. Go find someone who's said they're infallible and argue with them, then. I acknowledged above that they aren't.
  7. Because I know that the computers were good enough for Major League Baseball and the umpires' union to agree that they be graded on the results of computer evaluations. If it's good enough for them, it's good enough for me. Incidentally, that computer grading / training system is working to make them more accurate: FiveThirtyEight: Umpires Are Less Blind Than They Used To Be On the one hand, it's great that umpires have gotten better with the PitchF/X (now Statcast) feedback on their ball/strike performance for each game. On the other hand.... the best home plate ump still missed ~10% of his calls in 2015, as defined by what the computer thinks the rulebook strike zone is for those pitches. You can believe that the system is wrong, and that the umps are really getting it right. Or you can accept what MLB and the umpires themselves do, which is that this tech is much better equipped for the task of calling balls and strikes than even the best human beings are.
  8. You ever hear that saying "you don't have to outrun the bear?" The tech doesn't have to be perfect -- it just has to be measurably better than home plate umps are now, and I have no doubt in my mind that is the case already. I'm far more sympathetic to arguments that the current inconsistency / quirkiness / "game within the game" that exists with umps, catchers, pitchers, and batters is worth preserving than I am to arguments that somehow the tech isn't going to be perfect. Of course it's not -- but human beings aren't either. Try 4+ inches. The excel dump shows plate_x as -1.0963, or over one foot from the center of the plate. The plate's 17 inches, divide by two means the inside edge is at -8.5 inches, or about -.7 feet. Subtract that from the plate_x and you get 0.38 feet, or ~4.6 inches. Except, as I pointed out above, the CF camera is offset in the direction that makes a ball on the outside of the plate to a left-handed batter look like more of a strike. Look, computers aren't infallible. I get that. But the tech has been tested and improved over the years to the point where I have very little doubt that it's going to call the rulebook zone more accurately than umpires can. I respect people who say "I know the umps make mistakes, and that's the way I want it." But to argue that the tech isn't already doing a better job identifying strikes is to deny what's right in front of our eyes. The umpires themselves are graded based on how they compare to the computers, and this was true even when the computer tech wasn't nearly as good as it is now. If it's good enough for MLB to help train the umps and make their judgement better, how can it not be better than the natural judgement of the umpires without any such training?
  9. Paul DeJong 2017 Outlook

    I'm guessing you don't actually hate saying it all that much.
  10. Nothing wrong with that. I get the attraction of all of those nuances, but I come down on the side of wanting first and foremost to have the game decided by the actions of the players, not mistakes / grudges / whatever of the officials. But in a game where the no-pitch intentional walk created controversy, I don't see them automating the zone anytime soon.
  11. It's only pointless if the only two possibilities are 0% and 100% decision making by the computer. One can remove some, but not all of the "human element" and still make the calls much better. My point was that humans can participate in the computer's decision-making process, which is what was done with Pitch F/x, and is likely what's still being done by Statcast for the changing values of sz_top and sz_bot. In the future, it's possible that image processing alone could determine the upper and lower bounds reliably enough that humans wouldn't be necessary, but I doubt we're close to that right now. I'm not sure where you're getting the idea that I'm "jumping up and down" or "outraged" about this. Bellinger had an RBI single on that play, so I was perfectly happy with the result. That just happens to be the game and batter I chose for my query because I was watching the game yesterday, and it, along with the Brooks plots, clearly demonstrate an ump with no command of the horizontal strike zone. As I said, if we assume for the sake of argument that that one pitch was a rulebook strike, then he missed two strike calls that were either the same horizontal location or closer. It's right there in the strike zone plot. This all makes your points about vertical zone secondary. If we can't trust them to call the zone right for a 17" wide horizontal space that never moves, we certainly can't trust them to do that while also getting the vertical boundaries right without the benefit of a head-on view from center field.
  12. I didn't jump to that conclusion -- it's been mentioned in many articles about the Pitch F/x system. You made many points, including this one. I didn't address this one because I don't disagree. The TV strike zones are just there as a guideline. I get that. Look at where the CF camera is -- offset to the right. Look at how the pitch breaks -- right-to-left. *Maybe* it comes out of his hand catching some corner, but the break is away from Bellinger, and when the pitch crosses the plate, it's clear when you watch the video that it catches none of the black of the plate. Here's the ump's zone vs. LHB: Check out the left side of that plot. If you're right that the pitch caught the corner, what about the green square right below it at the same horizontal position? What about the green square that's more of a strike, but was called a ball? These inconsistencies happen in every game, because humans make mistakes. Computers can't solve everything, but they would consistently do a better job than humans alone.
  13. It actually is. Go to Baseball Savant's Statcast Search and look up any player's data for a single game. If you download the CSV data, you'll see that the sz_top and sz_bot values change pitch-to-pitch. In all likelihood, this is done the same way as it was with Pitch F/x, with a human analyst using the CF camera to judge where the rulebook zone boundaries should be for that batter, but unlike Pitch F/x, we see variations in the Statcast sz_top/sz_bot values pitch-to-pitch, which suggests they are actually trying to measure the rulebook strike zone on each pitch. Here's Cody Bellinger's pitch data from yesterday: So it's false to say Statcast isn't considering batter stance on each pitch for calculating the zone. Now, analysts did find some problems with Pitch F/x's data for sz_top and sz_bot many years ago, but I'm not aware of a more recent analysis, or any analysis of the newer Statcast system. The question is whether you trust an umpire to make this judgement about the rulebook strike zone better than the Statcast analysts while they're also trying to follow pitch location. That analysis uncovers some significant problems with umpire judgements -- e.g. four inches of increased height only leads to a strike zone that's 0.3 inches higher. At the end of the day, if you're using a human analyst to judge the top/bottom of the zone from a head-on view of the batter, I think they'll do a much better job than the ump will standing behind him and the catcher trying to call balls and strikes. As an aside: I picked Bellinger from yesterday for this exercise just to see what the data said about his 2nd AB yesterday. I was watching and one of the pitches looked waaay outside. Here it is on Brooks: And here's what the pitch looked like: The Statcast data had this as a ball ("zone" field 13 -- zones 1-9 are in the rulebook zone). This has nothing to do with the high/low strike, but just thought it was a pretty egregious call that makes a strong case for automation. I know this is just one bad call, but if the ump can't get that one right -- the plate doesn't get wider on each pitch, after all! -- how are we to trust him to be accurately judging where the midpoint of the batter's torso is on each pitch?
  14. Pro tennis has had line calls automated or instantly reviewable, and the equipment is rarely a problem. But if there is a total failure, there's still going to be a home plate ump -- they have to be there for purposes other than calling balls/strikes. They'll still watch the pitches, and can still be there to call balls/strikes in the rare instance of equipment failure. This is, at the very least, debatable: I thought from the beginning that the 1-2 WAR that people were attributing to framing was overstating the case, and the fact that framing isn't consistent year to year, and that the gap between the best and worst framers is narrowing, suggests it's not as valuable a skill as initially thought, because it involves fooling umpires, and they've made adjustments. There are also a lot of other catching skills that might not all go into a WAR calculation, but are nonetheless very important. Calling a game, working with pitchers on pitch sequencing / approach to getting hitters out, fielding wild pitches / balls in the dirt, controlling the running game, and normal fielding stuff like fielding pop-ups, plays at the plate, etc. Not having to worry about presenting the pitch isn't going to suddenly destroy the catcher position.
  15. Cody Bellinger 2017 Outlook

    Dude was born in 1995. Seinfeld ended when he was 3. Yeah, it's all over TV in syndication, but I'm guessing a baseball prodigy didn't spend much time flipping through the channels to watch old TV shows. Re: ADP, I agree that he's off the board toward the mid/end of the 2nd.