Page 1 of 1

Update Stats to a SABER-Analytics level in Mystery Leagues

PostPosted: Wed Sep 16, 2020 4:04 pm
by HoClarke
SOM mystery stats don't go as deep as a manager needs to make a line-up/pitching rotation decision.

Simple example: we get L/R - H/R stats, but surely a manger needs to know how a batter bats against a lefty H/R and against a righty H/R. That would be 3-Dimensional, but right now we only get 2-D. Same for pitchers.

And drilling deeper: a stat tool like BBREF where you can call up what a batter has done in played parks against a specific pitcher and vice-versa.

Assume a significant coding task, but also assume SOM is collecting these stats from game box-scores anyway, just not publishing them .

Would make for a more robust experience and not force a manger into blind decisions.

Such as: I see Johnny Temple hits lefties better than righties, and JT hits at home better than on the road, but does JT always hit lefties better at home, or does he maybe hit lefties better on the road?

And how does JT fare against the 3 scheduled pitchers today against him? (H/R)

Simple stuff in the Analytics Age. Not asking for grass/turf, night/day deep diving. Just something obvious. (But guess that also takes some of the mystery away...)

Re: Update Stats to a SABER-Analytics level in Mystery Leagu

PostPosted: Wed Sep 16, 2020 6:36 pm
by coyote303
It's interesting what you are asking for, but I don't understand how it can help. After all, we are talking about virtual pieces of cardboard. Player John Doe's card is the same regardless who it pitching, and his chance to get a hit is increased or decreased by a set amount (approximately .010 batting average difference) depending if he is home or away. Our fictional John Doe doesn't have any pitcher's number, nor does any pitcher have a special talent at getting Doe out. Any generated statistics suggesting otherwise in our simulated seasons are simply random chance.

So, I am curious how you would use this added information?

Re: Update Stats to a SABER-Analytics level in Mystery Leagu

PostPosted: Mon Sep 21, 2020 9:57 pm
by HoClarke
coyote303 - thanks for the reply. If it's all simply random chance, then I might be mistaken. Maybe that's why I lose so much!

I do make line-up / pitching choices based on the L/R, Home/Road stats provided, assuming outcomes are influenced by the park played in. In the mystery leagues most managers seem to pick a park and build a team around the park tendencies, as: in a park favoring right-handed home runs, draft right-handed home run hitters while drafting right-handed pitchers who don't give up many homers.

I guess my Johnny Temple example didn't explain/provide enough context?

I'm thinking of when I play in '61 Yankee Stadium, which the game states favoring lefties. Temple is a righty with all cards favoring him against lefties, so naturally I start him against a left-handed pitcher. But in Yankee Stadium, a righty has diminished statistical chances as listed in the Park card. So I'd like to know when Temple hits at home, is he successful against lefties or righties or nobody? The stats provided may state he hits .250 at home, and his L/R split states .275/.225, but I would still like to know how he specifically hits against lefties in Yankee Stadium. Altho unlikely, I want to make sure he's not actually hitting better against righties at Yankee Stadium, and I have no way of confirming this with the current stat reporting.

I surely assume he's hitting lefties at home better than righties, but I'd like to be sure. (and yes, OPS is really the determining factor, just using .avg to simplify). But if Temple was not doing that in games already played, this is where you could use a bench player who has shown better results in the specific L/R - H/R situation. I change my line-ups every series, depending on Home-Road stats currently reported, so looking for some confirmation of my decisions.

But besides the SABR-stat accuracy, I also think it's makes the game more fun to play and allows a manager to become more engaged with the team being managed when you have as many data points as possible. Baseball is not always about winning, it's also about enjoying the stats and SOM stat-reporting seems minimal in this day and age of analytics. The box-scores seem stuck in the 60s.

Again, if it really is all random - a lot of fun is taken out of managing anyway, but deeper stats would contribute to the illusion of fun for at least 2/3 of a season.


Re: Update Stats to a SABER-Analytics level in Mystery Leagu

PostPosted: Wed Sep 23, 2020 2:03 am
by coyote303
I understand better now what you're trying to achieve. However, I caution against using too small of a sample size to determine your strategy. Basing decisions solely on results is a dangerous game. In fact, I recently made a post about managers jumping the gun too soon on personnel decisions based on too few at bats in the Mystery League subforum.

I'll give you an example. Norm Larker (60s) was hitting up a storm. I analyzed each season's card and looked at doubles, homeruns, walks, as well as average. Everything pointed to his best card. He had well over 100 at bats. So, I traded my starting first baseman so I could use Larker full time (against righties). A few series later, he gets his first injury which proved he doesn't have his best card.

So, if you want to base using Temple (or a sub) after a couple of series in Yankee Stadium, there simply won't be enough data to reach a reliable conclusion.

This isn't to say you shouldn't be adjusting your lineup based on known data. With my current team, I sometimes make a change (or rarely two changes) based on an opposing pitcher's left/right results. With some teams I don't bother because all my starters are simply much better than their replacement even taking a less favorable matchup into account.

Anyway, I applaud you for doing your homework and trying to get the most out of your cards/players. Micromanaging makes the mystery leagues more fun! Just make sure your adjustments are for the right reasons and not just because someone had a poor (or great) series or two in a particular park.

Re: Update Stats to a SABER-Analytics level in Mystery Leagu

PostPosted: Thu Sep 24, 2020 5:49 pm
by eelbrain
coyote303 - Agree on not making decisions based on limited data. I've made choices similar to your Larker's and regretted.

Guess the simple point is: The more data the better! The more fun to play SOM!

Again, I assert: SOM is not providing deep-dive analytical data as much as it could. And I don't see why not. Maybe some/lots of managers wouldn't use it, but what's the harm in giving detailed statistical data?

Why not update to the 21st century and give more enjoyment to the game?