The Busted Toy?

[Max Power]

I was just thumbing through the STATS Inc 1992 Baseball Scoreboard - - the essay on BJ's Fav Toy - - - - when published then, it said that:

Ruben Sierra (26.9%) and Julio Franco (25.3%) had a better chance to get 3000 hits than Wade Boggs (17.1%) and Tony Gwynn (15.9%). And, that

Jose Canseco (44%) and Darryl Starwberry (34%) had a better chance at 500 HRs than Fred McGriff (17%) and Mark McGwire (13%).

What does this say about the Favorite Toy - - even in this small sample?

[Rajah]

Originally posted by Max Power
I was just thumbing through the STATS Inc 1992 Baseball Scoreboard - - the essay on BJ's Fav Toy - - - - when published then, it said that:

Ruben Sierra (26.9%) and Julio Franco (25.3%) had a better chance to get 3000 hits than Wade Boggs (17.1%) and Tony Gwynn (15.9%). And, that

Jose Canseco (44%) and Darryl Starwberry (34%) had a better chance at 500 HRs than Fred McGriff (17%) and Mark McGwire (13%).

What does this say about the Favorite Toy - - even in this small sample?

A lot of things that one simply can't predict skewed the results. Franco bounced around between the Mexican Leagues and the US, being a part time player in the US (and he's very likely older than we think he is, but BJ didn't know Julio's true age). Jose could still get his 500 if someone gives him a shot. Strawberry got involved in the drugs, and how could BJ have predicted the effects upon his career. McGwire was way injury prone, and BJ couldn't have predicted his relatively good health upon arrival in St Lous. When it comes to the others, I don't have much to say, due to lack of information.

[sweaver]

At that point, Sierra and Franco WERE more likely to get 3000 hits than Boggs and Gwynn. The same thing with the homers. Like Joaquin Andujar said, youneverknow. That's why they play the games.

[pwdennis]

I always wondered why nothing more was mentioned about "Favorite Toy" after a certain point. In one of his books James ran a projection on Wade Boggs' career - it would have been interesting if he had done a range of such projections for other players at that same point in time.

[Dr. Memory]

Originally posted by Max Power
What does this say about the Favorite Toy - - even in this small sample?

I believe the percentages were well founded. From that I suppose that the confidence interval was probably quite wide. In other words, the operative word was "toy". At any rate, James was clearly on to himself. It was a fun thing.

In re McGwire, he had 42 in '92, but injuries (which by the way he was not known for in 1992) caused him to hit only 18 HR total in 1993-94, so it was looking pretty good. I don't think anyone saw the coming league-wide HR explosion. Big Mac hit 245 in 1996-99, which was wholly unprecedented.

[KCBOOMER]

Originally posted by Dr. Memory
I believe the percentages were well founded. From that I suppose that the confidence interval was probably quite wide. In other words, the operative word was "toy".

I suspect the fact these projections were expressed as probabilities means they were just that. Informed hunches, if you will. I suspect age is the driving factor in these types of prjoections. For example, who is going to hit the most career homeruns, the guy who has 69 by his 23rd birthday (Canseco) or the guy who has 3 (McGwire)?

[SmedIndy]

It was an informed SWAG, nothing more. I always thought it was something to play around with - but stat-haters always jump on things like that to blast at James' credibility. Like THEY knew...

[Max Power]

Originally posted by SmedIndy
It was an informed SWAG, nothing more. I always thought it was something to play around with - but stat-haters always jump on things like that to blast at James' credibility. Like THEY knew...

Smed, after 1992, Boggs had 2,098 hits and Sierra had 1,160. And, Sierra had 10% more of a chance at 3000 than Boggs?

Believe, I knew, I knew............ ;)

[gyb13]

boggs was 7 years older. for sierra to reach boggs' hit number, he would have to average only 134 hits/season over the following seven seasons. By the end of 1992, Sierra was only 26 and had averaged 166 hits/season over his first seven seasons in the bigs.

[mandamin]

Originally posted by Max Power
Smed, after 1992, Boggs had 2,098 hits and Sierra had 1,160. And, Sierra had 10% more of a chance at 3000 than Boggs?

Believe, I knew, I knew............ ;)

They are also 9 years apart in age, and Sierra had hit .300 every year in his career, and walked less often. Boggs was coming off what would be the worst year of his career, bar none, hitting .259. When you're a career .335 hitter (and a very consistent one) and you hit .259 at age 34, you're looking pretty done. I don't think it was that unrealistic a prediction.

[Max Power]

I realize the age difference - and the walk rates.

But, I also know/knew, that, before you get to 3,000, you need 2,000. And, lots of yound hackers get to 1,000 and never make it close to 3,000.............it's a long road.

[mandamin]

Originally posted by Max Power
I realize the age difference - and the walk rates.

But, I also know/knew, that, before you get to 3,000, you need 2,000. And, lots of yound hackers get to 1,000 and never make it close to 3,000.............it's a long road.
Agreed. But, there aren't a lot of hackers that make it to 1000 by age 25, hitting .300 with good power numbers all the way. Of those that do, I bet that something close to one in every four of them eventually gets up to around 3,000.

Meanwhile, if you've got a 34 year old with 2,098 hits, you have to figure he has five good years left. Boggs is a great hitter, but he just had a horrible year, he walks a lot, and you have to figure he'll get less and less playing time, so 180 hits a year for the next five years willl be hard to do. Obviously, we know what really happened, but if I'm standing at the end of 1992, I'm probably thinking along those lines. I'm trying to think of a good comp to today--can't think of a match for Boggs (Frank Thomas comes close, great hitter in his mid 30's who seems to be fading), but Vlad would probably be a good comp for Sierra at similar ages. Who looks more likely to get to those milestones right now, Vlad or the Big Hurt?

And it's not like either was a sure thing. They were both long shots, and the 17% happened to hit while the 26% didn't. Not entirely out of the ordinary. Now, if you ran Boggs, Gwynn, and Ripken alongside a bunch of guys that didn't come close and came out with consistently way-off numbers like that, then I'd say there's a problem with the system. I just think that the few little samples you've given aren't very conclusive.

[SmedIndy]

Originally posted by Max Power
I realize the age difference - and the walk rates.

But, I also know/knew, that, before you get to 3,000, you need 2,000. And, lots of yound hackers get to 1,000 and never make it close to 3,000.............it's a long road.

It's also true that a lot of guys Boggs' age in 1992 just STOPPED hitting....it's ludicrous only in hindsight.

[Max Power]

How many of these guys made 3,000 hits?


CAREER
AGE < 26
HITS > 900
AGE <= 25
SLG > .470

HITS H H SLG
1 Ty Cobb 1433 1433 .514
2 Mel Ott 1249 1249 .554
3 Al Kaline 1200 1200 .480
4 Freddy Lindstrom 1186 1186 .471
5 Vada Pinson 1177 1177 .485
6 Alex Rodriguez 1167 1167 .571
7 Hank Aaron 1137 1137 .559
8 Jimmie Foxx 1127 1127 .638
9 Orlando Cepeda 1105 1105 .537
10 Joe Medwick 1101 1101 .564
11 Mickey Mantle 1080 1080 .574
12 Rogers Hornsby 1073 1073 .499
13 Arky Vaughan 1057 1057 .491
14 Ken Griffey Jr. 1039 1039 .536
15 Frank Robinson 994 994 .561
16 Ruben Sierra 993 993 .474
17 Joe Jackson 988 988 .528
18 Joe Kelley 980 980 .511
T19 Joe DiMaggio 970 970 .623
T19 Hal Trosky 970 970 .559
21 Tris Speaker 958 958 .492
22 Andruw Jones 940 940 .491
23 Ron Santo 933 933 .471
24 Willie Keeler 931 931 .506
25 Cal Ripken 927 927 .483
26 Johnny Bench 922 922 .479
27 Goose Goslin 910 910 .503
28 Eddie Mathews 902 902 .556

Maybe a third?

[JamesI]

I wouldn't consider the toy broken, nor is it a great representation of the future. I've never heard of the toy giving good odds of anything.

I am curious as to why Gwynn was given such a poor chance of reaching 3000. Boggs I understand since it looked like he was losing his ability (and at that point he was not the same player he had been).

[mandamin]

Originally posted by Max Power
How many of these guys made 3,000 hits?

Maybe a third?

No, by my count, it was nowhere near a third (I could be wrong...I'm bad at keeping track of who has which milestones :)). If it is anywhere near a third, of course, that makes my point for me. I have to admit the 26% might be a touch too high, though, although A-Rod and to a lesser extent Andruw still have a chance, and some of these guys lost time to war or other oddball considerations (like being banned from the sport). Also, I wonder what you'd get if you ran a list of guys with around 2000 hits at age 34.

[poorme]

well, what is the history of 34 year olds with 2100 hits?

oops...is there an echo in here?

[KCBOOMER]

Max, i trying to find a point here. All statistical projections fail occasionally and that is all it is. To say the theory is wrong because a guy who is projected as having a 35% chance to achieve something (i.e. a 65% chance of not achieving it) and doesn't while a guy who only has a projected 25% chance does doesn't invalidate the theory.

If we are going to do that the California Angels just destroyed the entire mantra of OBA and the value of walks.

[Max Power]

Originally posted by poorme
well, what is the history of 34 year olds with 2100 hits?


HITS >= 2100
AGE <= 34
SLG <= .470

HITS H H SLG
1 Robin Yount 2747 2747 .437
2 Sam Crawford 2668 2668 .458
3 Willie Keeler 2650 2650 .434
4 Vada Pinson 2574 2574 .449
5 Roberto Clemente 2559 2559 .469
T6 Jesse Burkett 2547 2547 .463
T6 Pete Rose 2547 2547 .432
8 Roberto Alomar 2546 2546 .450
9 Frankie Frisch 2518 2518 .443
10 Richie Ashburn 2455 2455 .382
11 Nellie Fox 2395 2395 .369
12 Rod Carew 2394 2394 .444
13 Eddie Collins 2379 2379 .426
14 Cal Ripken 2371 2371 .453
15 Stuffy McInnis 2368 2368 .381
16 Rusty Staub 2364 2364 .433
17 Pie Traynor 2357 2357 .437
18 Jake Beckley 2351 2351 .449
19 George Davis 2315 2315 .426
20 Lloyd Waner 2306 2306 .400
21 Buddy Bell 2273 2273 .409
22 Willie Davis 2271 2271 .414
23 Brooks Robinson 2259 2259 .421
24 Hugh Duffy 2257 2257 .450
25 George Van Haltren 2255 2255 .424
26 Ron Santo 2254 2254 .464
27 Joe Torre 2238 2238 .455
T28 Joe Sewell 2226 2226 .413
T28 Ted Simmons 2226 2226 .443
30 Billy Herman 2205 2205 .407
31 Tony Gwynn 2204 2204 .446
32 Red Schoendienst 2200 2200 .393
33 Lou Brock 2194 2194 .425
34 Charlie Grimm 2187 2187 .400
35 Sherry Magee 2169 2169 .427
36 Joe Kelley 2161 2161 .455
T37 Max Carey 2158 2158 .388
T37 Al Oliver 2158 2158 .458
39 Tim Raines 2152 2152 .428
40 Rickey Henderson 2139 2139 .443
41 Ryne Sandberg 2133 2133 .455
42 Fred Clarke 2126 2126 .439
43 Zack Wheat 2121 2121 .435
44 Keith Hernandez 2106 2106 .443
45 Monte Ward 2105 2105 .341

[Max Power]

Originally posted by KCBOOMER
Max, i trying to find a point here.

Actually, just making conversation. Never meant to make a case - just asked a question, really.

[mandamin]

Ah, there we have it. If you look toward the bottom of that list, which is the guys you could group Boggs with (although actually, you could pretty much go up to like #15 on the list with the same results), very few of those guys actually made it to 3000 hits. So I can see how if you were devising a system, it could defensibly pick Sierra as the better candidate than Boggs.

Of course, you could then look at it subjectively and say "well, Boggs is a brilliant, professional and talented hitter who is healthy and just had eye surgery (didn't he?) and never hit below .300 until this year, so he'd be a good candidate to make a run for it," and then say "well, Sierra is a hacker with a huge hole in his swing and he seems kind of lazy. So I think I'd flip those two." But since most 34 year olds with 2100 hits are not Wade Boggs and a lot of 25 year olds with 1000 hits are not Ruben Sierra, I think it's perfectly understandable for the Favorite Toy to call it the way it did.

It's really fun to look back on those things, though. Like his projections in "Whatever Happened to the Hall of Fame" of who would be elected to the Hall and when. Sierra is a notable goof-up on that one, too. :)

[Ytown Tribe fan]

BJ isn't alone.

Won't we ALL look silly when NO ONE breaks Aaron's HR record -- EVER?

[Fuzzy Bear]

Originally posted by JamesI
I wouldn't consider the toy broken, nor is it a great representation of the future. I've never heard of the toy giving good odds of anything.

I am curious as to why Gwynn was given such a poor chance of reaching 3000. Boggs I understand since it looked like he was losing his ability (and at that point he was not the same player he had been).

One reason Gwynn was given low odds at reaching 3,000 hits was that at age 24 he had only attained 149 hits and a .296 BA (in two partial seasons prior). Gwynn got a late start; it is much harder to attain 3,000 hits starting late unless you are an extreme high-average hitter (like Boggs and Gwynn).

Another reason that Gwynn received low odds is the comparable length of careers to date in 1992. In 1992, according to what had transpired before then, many of the greats retired before age 40. (Mickey Mantle was done after age 37, and this was not unusual.) In the past 10 years, career length of the great players has gone up significantly. If BJ used the favorite toy on Tony Gwynn today, assuming Gwynn was as far into his career as he was in 1992, and EVERYTHING ELSE BEING AS IT IS NOW, I suspect that Gwynn's chances at 3,000 hits would be higher.

As for Sierra, no one foresaw that he would ruin a HOF career with an ill-advised weight-training (steroid?) program that would completely change the type of player he was. The Favorite Toy is not equipped to handle that any more than Strawberry's drug problems.

[Rinkster]

Sheesh folks, James called his method a "toy", a FREAKIN' TOY!

Think there was a reason he called his method The Favorite Toy rather than, say, The Career Prediction Commandments?

It's supposed a fun formula to play around with when you have an hour or two to kill. The fun part is that you are GUESSING at what a player should be capable of doing over the course of his career, and enjoy watching how a particular player surpassed what should have been expected of him--or to see how circumstances impossible to predict (injuries, trades, success, drugs, weight-training, MARRIAGE) cause a player to come up short of his "possible" capabilities.

Like it or not, the system works pretty well in cases where nothing major pops up and derails a career.

[Ytown Tribe fan]

You tell 'em, Rinkster.

No, seriously, you tell him next time you see him. :D

[Fuzzy Bear]

Another thing about the Favorite Toy is that, as a player gets older, a player's chances to attain a milestone go up or down vastly.

For example, after the 1999 season, Harold Baines' chances of getting 3,000 hits and 400 homers skyrocketed. Yet after the 2000 season, his chances plummeted for both goals, and were virtually extinguished after the 2001 season.

The Favorite Toy probably showed Baines' chances for 3,000 hits and 400 homers as better than even after 1999, about 25% or less after 2000 and probably in single digits now after the 2001 season (assuming Baines would play in 2002, which he did not).

This is not a malfunction of the Favorite Toy. It merely reflects the fact that an older player is often one good season away from being less than a full season away from a milestone, but one injury or slump away from the end of his career.

[pwdennis]

Originally posted by Fuzzy Bear
Another thing about the Favorite Toy is that, as a player gets older, a player's chances to attain a milestone go up or down vastly....This is not a malfunction of the Favorite Toy. It merely reflects the fact that an older player is often one good season away from being less than a full season away from a milestone, but one injury or slump away from the end of his career.

Good point. It would be interesting to see Favorite Toy projections on the last few seasons of Mark McGwire (or before him Jackie Jenson who was only a year removed from being the AL MVP when he retired). Of course, the Toy doesn't (and can't/shouldn't) project for unexpected retirements Jenson, Ryno or Big Mac.

[Fuzzy Bear]

Originally posted by pwdennis
It would be interesting to see Favorite Toy projections on the last few seasons of Mark McGwire (or before him Jackie Jenson who was only a year removed from being the AL MVP when he retired). Of course, the Toy doesn't (and can't/shouldn't) project for unexpected retirements Jenson, Ryno or Big Mac.

The toy may well have discounted Big Mac's chances to reach 600 after the 2001 season due to his radical dropoff in ABs (less than 300 two years in a row) and his sub-.200 BA in 2001.

I wrote an essay on another website predicting that McGwire would retire after the 2001 season. I had a hunch that he had other things on his mind, and that his knee injury was of the chronic variety. I was a little surprised when I was right, but, in hindsight, one could see that coming. There are a surprising number of players who have career years in their late 30s that are done 2-3 years after it. This may be of interest to Mr. Bonds and Mr. Aaron.