[Pennant Chase] A change is coming... to Arm Rating!

[GM Games 252]

Arm rating. One of the most despised and questioned aspects of Pennant Chase. Let’s quickly dive into the history: The arm rating applies only to catchers and outfielders. It was a way to determine which players would be more likely to 1) throw out base stealers or 2) throw out a runner on the bases from the outfield. Infielders do not have an arm rating, because the intent of the sim is that the ball in play is either an out or it’s not. (More to come next time when we dive into Range Factor.)When PC originally launched, we didn’t have any stats to determine the arm rating, thus it was mainly available for custom leagues who wanted to manually distinguish certain players. The arm rating is binary: it’s either a normal arm, or a plus arm. Plus arms at the catcher position are particularly valuable because they can have a true impact on the opposing team’s running game. So much so that users suggested I implement a manager’s option to shut down the run game when facing a plus-arm catcher. That was put into place in the early years of the website.As a quick-and-dirty solution, we used a player’s assist totals to determine his arm rating. (As most of you know, all aspects of PC are driven by real-life stats. There are legal reasons for this, but it’s also the tenant of the website to not have any arbitrary “ratings”.) For outfielders, that works okay. If you had a lot of assists in the outfield that season, you deserve a plus arm so that it can be reflected in the sim. It’s not perfect. Every strong-armed outfielder has had seasons in which he didn’t throw many people out. In fact, he may even get fewer opportunities because of his strong arm. But that’s okay, because this is a stats-driven sim. If he didn’t throw many people out in real life, why should he throw people out in the sim just because he has a reputation for having a strong arm? The issue is that people who don’t understand this get confused by the arm rating.The other small problem is that players who split time in the outfield and infield have a boatload of assists and thus automatically get the plus arm. But again, this is a minor problem. For one, if they are an infielder they probably deserve the plus arm in the outfield. Secondly, if outfield is their secondary position, the risks far outweigh the benefit of the plus arm.But for catchers, it’s been massively problematic. In the past, we didn’t have caught-stealing stats in the database. After a lot of grueling work, we now do!! So let’s analyze how this could impact arm rating.Let’s look at Buster Posey in the 2017 Auto Leagues. He has a plus arm because he had 52 assists as a catcher and 25 assists as a first baseman. But his caught stealing percentage in 2017 was 38%. That’s basically average in baseball history. If you look only at catchers with at least 30 attempts, the median in baseball history is 37%. Posey was perfectly average at throwing out base runners. But his plus arm is leading to a 47% rate across 2017 Auto Leagues.The best catcher in 2017 was Wellington Castillo of the Orioles, who threw out almost 50% of attempts against him in 50 tries. But he does not have a plus arm in Pennant Chase. As a result, he’s throwing out just 39% of baserunners in 2017 leagues. It’s pretty clear this should be exactly reversed. Posey should have an average arm, Castillo should have a plus arm. Let’s take a look at one more fascinating example. Wilson Contreras of the Chicago Cubs is an athletic catcher who is known for having an arm. Sure enough, he has a plus arm in 2017 leagues. But that’s only because he had 86 assists in 2017. Yes, 86!! And get this - 36 of those assists were due to dropped third strikes resulting in the batter being thrown out at first. Holy moly, either Contreras sucks at catching the ball, or his pitchers love throwing the ball in the dirt. Would you be surprised to know Contreras only threw out 27% of base stealers? In the context of throwing out base runners, Contreras should definitely not be a plus arm. Now, there are counter arguments: 1) His pitchers could have been terrible at holding on runners. True, but unfortunately, we’re never going to know that, we have to go with the stats we have. 2) His high assist total does demonstrate that he has a good arm. He obviously threw a lot of runners out in various situations even if they were not stolen base attempts. Also true, but that benefit can be covered by the sim in a different way (as I said, stay tuned for the Range Factor discussion). If we are only looking at stealing bases, Wilson Contreras wasn’t that good at throwing out base stealers in 2017.Now the big announcementSo what is the plan? Well, I feel the arm rating works pretty well in the sim. I thought long and hard about switching to a hard-line caught-stealing percentage, but that doesn’t make a ton of sense. First off, the sample size is really small for a lot of players. If a catcher threw out two of three base stealers, should he really get credited for a 67% success rate? That would destroy the sim. Also, if we were to use percentages, we would need to balance the catcher’s rate with the base runner’s success rate. At this time, I don’t have baserunning stats in the database, so that’s not an option.What I think works the best is changing catcher’s arm to be based on their success rate. If a catcher had at least 30 attempts and threw out at least 40% of base runners, that would put him in the top 40% of catchers in MLB history. That feels like a good number to consider a plus arm. (We can always tweak it.)As an example, what would this do for a player like Ivan Rodriguez, who we know had one of the best arms in history? Well, 12 of his 20 qualifying seasons would result in a plus arm - that’s four more plus-arm seasons than the current formula, which is based on assists. In addition, 10 of his first 11 seasons would be plus-arm, which makes sense considering a younger I-Rod would have a better arm than an older I-Rod. So - this change is something that has to happen across all leagues Auto Leagues at once. (The date is TBD, but I’ll announce when it happens on the Public Boards.) That’s always hard because it impacts seasons in-flight. People who had a plus-arm catcher suddenly are not going to have one. I know that sucks, but changes like this that are good for the long-term accuracy of the site, people tend to be okay with. Custom leagues are a little different since they could have used progression or fictional players. I don’t want to impact custom leagues, so I’m going to request commissioners reach out to me if they want me to make the change in their league. Moving forward, any player imports from the historical database will reflect this change.

 

View the full article