SLB Cheating Mike LaValliere's Arm at Catcher? - SimLeague Baseball

Posts: 2106 (0)

Hall of Famer

266 1 37

Prospect

Witness the strange case of a man with ostensibly an outstanding partial season behind the plate -- in the real world, but not in the SimLeage Baseball Zone.

Real Life, according to Baseball-Reference.com, the year is 1993 and in 37 games behind the plate for the Chicago White Sox:

8 stolen bases allowed
24 base runners caught stealing
75% of the time, he throws them out!

SLB's record of Open League experience, on average:

16 SBA
25 attempts (that part is cool)
37%

#WhatGives?

6/15/2015 12:06 AM (edited)

crazystengel

Posts: 12465 (7)

Hall of Famer

714

Hall of Famer

In the draft center, 1993 Lavalliere's CS% shows as just 44%. Why isn't it 75%? No idea.

Hard to say what WIS does with CS%. To begin with, I think they cap it at 60%, regardless of real life numbers. So if you search by CS% in the draft center you'll see a bunch of guys with 60% on the nose -- with the exception of 3 guys who are over 60%, who are all part-timers from 2014, and who (presumably) somehow slipped through the cracks.

The best one is 2014 Stephen Vogt, at 95%. He's already been used 12 times in the sim, but his average CS% in those 12 seasons is only 46%. Of course, you have to take into account that most of the runners attempting steals in OLs (which is where Performance History stats come from) are Vince Coleman and Tim Raines. But still.

Then there's the whole issue of fatigue -- even serious fatigue (0%) -- barely impacting a catcher's throwing arm. And the fact that any catcher with a good arm but few PA is going to cost as much per PA as Babe Ruth in his prime. 2004 Pat Borders, for example (OPS# of .462) is the 23rd most expensive batter per PA in the entire sim. More expensive than 41 Williams, 24 Hornsby, 01 Lajoie, 11 Cobb, etc. Why? 55 PA of horrible hitting and an A+ arm? I don't get it.

6/15/2015 12:46 AM

uncleal

Posts: 1619 (0)

Hall of Famer

339 6 2 4

Rookie

CS% will be low in leagues with experienced owners because they won't run with runners that aren't among the most successful in the game, especially against good arms.

6/15/2015 1:29 AM

elbirdo

Posts: 129 (0)

Hall of Famer

877

Rookie

I think sim has wrong data for this. Draft profile shows 0 SBA and 0 CS.

6/15/2015 1:45 AM

contrarian23

Posts: 23482 (15)

Hall of Famer

1928 1

Hall of Famer

At a guess, there are a few things going on here.

1.) I am sure WIS discounts CS% for catchers who were part-timers. Which makes sense. You certainly don't want a guy who threw out 1 runner in 1 attempt to have a 100% CS as a regular catcher. They may give the guy an A+ arm, but the underlying percentage they use is almost certainly adjusted (on a related note, the A+ arm for Bert Campaneris in 1965 is an error. Campy famously played all 9 positions in 1 game that season, and did not throw out any runners in his one inning as a catcher. But he did have 1 assist in that game while playing other positions, so he "looks" like a strong-armed catcher from the perspective of assists/game.)
2.) More to the point, the guys running on Lavalliere weren't exactly a whos-who of basestealing threats from the era. Among the players he "caught" stealing were Bret Boone (2 SB in 5 attempts that season), Dean Palmer (11/21), Greg Gagne (10/22), Greg Litton (0/1), Juan Bell (6/13), Omar Vizquel (12/26), Reggie Jefferson (1/4). Whereas the guys who were successful were, as a rule, better basestealers: Albert Belle (23/35), Chad Curtis (48/72), Ruben Sierra (25/30), David Hulse (29/38), Kirk Gibson (14/21). I'm cherrypicking some data here - there were certainly a few counterexamples - but the point is it would not be realistic to take his 75% CS rate against a group of guys who were not, as a rule, good SB men...and have that translate into 75% throwing out Vince Coleman, Tim Raines, et al.
3.) Also on the small sample size point, at least 2 of his CS (could have been more, I didn't look carefully) were players who were thrown out at home as part of a doublesteal. The catcher gets an assist on that play if he throws the ball to (say) the SS who then returns it to home, but it's not clear that he deserves to be credited with a CS. In a full season, that stuff washes out, but in a small number of games it can make you look better than you are.
4.) Lavalliere in his career never had another performance like that. Again, small sample sizes.

All of which is to say, I am comfortable with downgrading him.

Add on to that the point that uncleal makes, which is that many owners set SB to 0 for anyone who was not a frequent and successful thief. And Sparky takes into account, in some fashion, the opposing catcher's arm before calling for a steal. So in WIS, Lavalliere probably only has the best of the best running against him. There's no way he's going to throw out the elite basestealers in MLB history at a 75% rate.

6/15/2015 2:03 AM

crazystengel

Posts: 12465 (7)

Hall of Famer

714

Hall of Famer

contrarian, you're no doubt correct that Lavalliere's 75% CS rate is not reflective of his "true" abilities, and that it's the result of a small sample size. However, those are the real life stats he accumulated, so shouldn't they be used? 1988 Bob Milacki's stats don't reflect his abilities, either, nor do 1998 Shane Spencer's, but WIS hasn't "downgraded" them at all. Wouldn't it be better to let these players have their actual stats and price them accordingly?

6/15/2015 2:59 AM

contrarian23

Posts: 23482 (15)

Hall of Famer

1928 1

Hall of Famer

crazystengel, you are of course correct that WIS's approach is inconsistent.

To the extent that they were to make a change in one approach or the other, I would much rather they downward-adjust Milacki, Spencer, et al rather than leave Lavalliere unchanged, but I am not holding my breath for either.

There is, though, I think one important distinction between the two cases, and this may be why WIS has 2 different approaches. My apologies if I am not particularly articulate here; working on very few hours of sleep before heading off to the stifling humidity of New Orleans.

Both examples (Milacki/Spencer and Lavalliere) are impacted by small sample size issues, and all of the randomness and variation that accompanies those samples. But with regard to stolen base attempts, there is an additional, decidedly less random, component. And that is that teams/managers/players can choose who attempts a steal and who doesn't. To a much greater degree than they can influence who hits against whom. Not every player, obviously, chooses to steal. Whereas everyone in the lineup has to hit. So exacerbating the small sample size issue is that a catcher's SB/CS data are highly influenced by the choices opposing teams/managers/players made in addition to the small sample size effects. And in the case of Lavalliere, quite a few of the SB attempts against him were by players who had very few such attempts on the season, and by other players who were disproportionately unsuccessful when they did try. It's not completely unreasonable for WIS to discount CS% in small sample sizes on that basis, even if they don't do so for other stats (like Spencers slash stats or Milacki's OAV.) I'm not saying I 100% agree, but they do at least have some logical basis for treating the two cases differently.

This is related to a larger issue I've been thinking about recently, and may find time to post about...and gets to the very nature of how simulations work (or should work).

6/15/2015 7:37 AM (edited)

ArlenWilliam

Posts: 2106 (0)

Hall of Famer

266 1 37

Prospect

Thanks, but not finding the explanations very satisfactory. I've seen plenty of partial season catchers with better results in SLB's OL stats. Maybe he didn't get enough games or innings to qualify, comparatively, so they give him average percentage. Maybe his name reminds them of one of the sets of "villains" in Stephen King fiction. God knows what rhyme or reason goes into some of their algorithms, bless 'em.

6/15/2015 12:48 PM (edited)

uncleal

Posts: 1619 (0)

Hall of Famer

339 6 2 4

Rookie

Posted by contrarian23 on 6/15/2015 7:37:00 AM (view original):
crazystengel, you are of course correct that WIS's approach is inconsistent.

To the extent that they were to make a change in one approach or the other, I would much rather they downward-adjust Milacki, Spencer, et al rather than leave Lavalliere unchanged, but I am not holding my breath for either.

There is, though, I think one important distinction between the two cases, and this may be why WIS has 2 different approaches. My apologies if I am not particularly articulate here; working on very few hours of sleep before heading off to the stifling humidity of New Orleans.

Both examples (Milacki/Spencer and Lavalliere) are impacted by small sample size issues, and all of the randomness and variation that accompanies those samples. But with regard to stolen base attempts, there is an additional, decidedly less random, component. And that is that teams/managers/players can choose who attempts a steal and who doesn't. To a much greater degree than they can influence who hits against whom. Not every player, obviously, chooses to steal. Whereas everyone in the lineup has to hit. So exacerbating the small sample size issue is that a catcher's SB/CS data are highly influenced by the choices opposing teams/managers/players made in addition to the small sample size effects. And in the case of Lavalliere, quite a few of the SB attempts against him were by players who had very few such attempts on the season, and by other players who were disproportionately unsuccessful when they did try. It's not completely unreasonable for WIS to discount CS% in small sample sizes on that basis, even if they don't do so for other stats (like Spencers slash stats or Milacki's OAV.) I'm not saying I 100% agree, but they do at least have some logical basis for treating the two cases differently.

This is related to a larger issue I've been thinking about recently, and may find time to post about...and gets to the very nature of how simulations work (or should work).

Well there's another major difference between CS% and, say, WHIP.

Bob Milacki's WHIP is going to start shooting up if you continue to throw him. A 0% Milacki sucks.

LaValliere's CS% isn't going anywhere even if he's playing at 0%. (Granted, he'll make errors and won't hit, but that's besides the point.) This is a mistake in the WIS algorithm, to be sure -- but it looks like they're sort of covering one mistake with another, possibly...

6/15/2015 12:55 PM

contrarian23

Posts: 23482 (15)

Hall of Famer

1928 1

Hall of Famer

I should also note that Lavalliere's performance history is based on 10 seasons of data. 10 seasons with few PA and a relatively small number of chances. I'd be careful about drawing any conclusions here.

6/16/2015 5:44 AM

ArlenWilliam

Posts: 2106 (0)

Hall of Famer

266 1 37

Prospect

Are we really to believe they either make iterative decisions like that about ballplayers -- or that they have logic so complex as to compare players' partial season stats to career stats for some purposes?

[EDIT: After reading your learned comments...] I don't have a guess but if I had to guess, I'd say he didn't qualify in WIS' minds with enough innings for his own rating (notwithstanding that the rating is capped) and they gave him an across the board-? 40% as an average of what A+ would be for a catcher's arm.

6/16/2015 12:31 PM (edited)

slashtc

Posts: 4745 (2)

Hall of Famer

305 185 1 183

Prospect

I'm of the opinion that the entire SB engine is screwed up. All you're seeing is just a part of that.

6/16/2015 12:24 PM

ArlenWilliam

Posts: 2106 (0)

Hall of Famer

266 1 37

Prospect

I think it's better to say it is a very elaborate an "heroic' task and they have always had significant bulges and creases to work on, simply by trying to statistically simulate baseball over one-hundred and thirty years, under one blanket.

Actually, about that, I suggest they create two versions of dead-ball era players, one for intra-era play and another for playing against players of other eras.

6/16/2015 12:33 PM

crazystengel

Posts: 12465 (7)

Hall of Famer

714

Hall of Famer

Posted by ArlenWilliam on 6/16/2015 12:31:00 PM (view original):
Are we really to believe they either make iterative decisions like that about ballplayers -- or that they have logic so complex as to compare players' partial season stats to career stats for some purposes?

[EDIT: After reading your learned comments...] I don't have a guess but if I had to guess, I'd say he didn't qualify in WIS' minds with enough innings for his own rating (notwithstanding that the rating is capped) and they gave him an across the board-? 40% as an average of what A+ would be for a catcher's arm.

I'd say he didn't qualify in WIS' minds with enough innings for his own rating

That's an interesting idea, one that I hadn't thought of. Maybe it's something like what they do for pitchers' batting stats -- like when you see a pitcher who went 1 for 2 in real life, he's not going to hit .500 for you in the sim. They "fill him out" with 48 more at-bats of a typical pitcher's batting stats.

6/16/2015 1:09 PM

ArlenWilliam

Posts: 2106 (0)

Hall of Famer

266 1 37

Prospect

Well[ put, CS. Now, should we discuss why modern ace relievers don't do well?

6/16/2015 11:59 PM

Forums

SLB Cheating Mike LaValliere's Arm at Catcher? Topic

Search Criteria