Sabremetrics/Traditional Opinions

[Bill Burgess]

We all look forward to the day when the stat record will confirm and illuminate the Traditional, Anecdotal Record of Contemporaries Opinions.

But until that day, we have many unresolved conflicts yet. Some toughies.

1. Stats declare Babe Ruth the greatest player ever. Virtually all who saw both play extensively, declared Babe the greatest slugger, but Cobb the greatest player. Unresolved.

2. Stats seem to indicate that Ted Williams was the second greatest player after Babe. All contemporaries always insisted that Ted was the better hitter, yet Joe DiMaggio was by far the greater all-around player, due to his far superior defense.

3. Stats seem to indicate that Mickey Mantle was the greater hitter, but contemporary opinion insisted that Willie Mays was the better player, due to his superior glove, and likely his running, even though Mickey ran very well.

So here's the poll/survey question.
If stats say that a hitter is the greater player, as in Williams/Musial, but traditional opinion insisted otherwise, who do you tend to lean towards - stats or opinion.

Can a hitter's lead, like Williams over DiMag/Musial be so over-whelming as to make up for all else.

[misterdirt]

I am old enough and lucky enough to have been able to see Mays and Mantle play in their prime and Williams at the end of his career. It gives me a thrill just to think about having been in the presence of such special players. My idea of heaven is to go back in time and spend endless lazy days at the ballpark watching the stars of the past like Ruth and Cobb play. My idea of hell is wasting time arguing over which of these great players was a little bit better than than the others.

[gotahavmypops2]

Everyone,
I always take stats over opinion. It's kind of like math. Some people say it is worthless but, the stats say that nearly every career uses math. You can't take a Negro from the leagues and say they were bad just because of word of mouth.

Ben Taylor is a good example of that. My uncle told me that Taylor was one of the worst baseball players ever. So I looked up the stats and found out how good of a player he actually was. I found out that he was one of the best first basemen the Negro Leagues have seen.

[jalbright]

I would say I take stats over opinion, but with an absolutely huge caveat: only if we can find a way to reasonably pull the biases out of the stats. Great examples are the major leagues of the twenties and thirties; players in the Astrodome versus the Baker Bowl, Coors Field, and Mile High Stadium and so on. If we don't do that, stats can be just as deceptive as opinion. It's best when the two come close to agreement, but if push comes to shove, I'll take carefully analyzed stats. It's easier for me to tell whether someone has carefully tried to take the distortions out of the statistical record. The very best subjective opinion probably trumps statistical information--but it's almost impossible to determine which opinion(s) to put that kind of faith in. Common wisdom has been shown to be wrong so often that it doesn't deserve that kind of faith.

Jim Albright

[SABR Matt]

OK

a) No I didn't see Ty Cobb play...if I did, I would not have NEARLY the brain pwoer I would need to make PCA worj given I'd be like 100.

b) Yes I take stats over opinions generally in deciphering the game...and here's why.

While it is true that the vision of people who knew a great deal about baseball did a pretty good job of picking out who should play and who should not...and it is true that although they didn't know about sabermetrics...they generally knew enough about the sport to maximize the ablities of the players under them...the good managers did at any rate...Rogers Hornsby...probably not .. .. .. baseball is a logical game...it flows from event to event in an extremely mathematically pure way. The game's statistics have evolved the way they have because peple even in the 1870s knew that run scoring was the product of the sum of a series of events following each other in (although physically complex, still logically simple) orderly fashion.

Because of this truth that 27 Outs = 1 game, 4 bases = 1 Run, and a win happens when you outscore your opponants by any degree...the statistics we develop today, while still works in progress can and will lead us to a lot more knowledge about how the game works than the working eyes of the people down on the field. I know that flies in the face of a lot of conventional wisdom on the subject, but that is what I believe.

[SABR Matt]

My last post also failed to mention a couple of known human biases.

1) Humans tend to prefer the more "well rounded" players...generally having a well rounded player is good, but on a baseball field, Sabermetricians are becoming more and more convinced that value is value is value, no matter how you achieve it.

To the naked eye, guys like Ty Cobb and Honus Wagner who do all of the little things right on both sides of the ball look like they're doing more for the club, but a single fielder in a defensive scheme can't have as much importance as the hitting componant...Honus Wagner was a 6 DW, 12 OW player in his prime...very good...well rounded player. But I have 9 people ahead of him on my all timelist because I found 9 players who were greater than 18 OW players in their prime regardless of defense. You don't want 9 players none-of-whom can field...but given the choice between a superstar hitter and a player who's very good in a lot of areas...I'll take the hitter.

2) People have a bad tendency to fondly remember the brilliant things a player did in his career "in the clutch" and forget some of the really dumb/bad/weak plays he made...or in the case of A-Rod...remember a few crucial strikeouts in the post-season and forget the incredible player he is day to day because of it.

Human observation is full of biases that no matter how carefully researched and no matter how many sources chosen...cannot be weeded out entirely. In a game that is logically pure and composed of linear combinations of events that lead to runs and outs that lead to games...I'll take the logical purity and common sense that comes from a careful study of sabermetrics.

[538280]

I think what it really boils down to is that the stats will always be the sure thing, you can almost never go wrong with them. The opinions really aren't a sure thing, and even the best of baseball men (Mack, McGraw) can sometimes be wrong, remember, no one is perfect.

[leecemark]

--The stats give us a good picture of the approximate value of a player. They aren't perfect, but for almost every player they are going to reflect something close to what they contributed to their teams. I start with the numbers and may adjust them a little up or down based on the subjective view comonly held of that player.
--If most baseball men thought a player was better or worse than their numbers, they probably were. Not much better or worse, most likely, but perhaps enough to jump them over another player they trail a little just by stats.

[CyNotSoYoung]

Well - I never saw either play, but here's my opinion. Ty was not better than Babe. Babe was not better than Ty. They were both excellent. They achieved the greatest heights of baseball excellence - but in different ways. And they were both the greatest. So was Willie Mays, so was Mickey Mantle, so was Walter Johnson. All were the greatest.

Do you drink wine? Just one vintage or do you choose one to suit how you feel that day? Do you eat food? What's the greatest thing you've ever eaten? Would you want to eat it every meal, every day? Do you like just one kind of music? Is there just one song that was the greatest ever?

You can rate things - food, wine, music, ballplayers till the cows come home. But in the end it won't make a bit of difference. To quote Duke Ellington - "If it sounds good, it is good."

[SABR Matt]

CyNotSoYoung:

You separate Stats from opinions by not starting with na agenda. I don't have an agenda...I don't start "trying to prove that Cobb is better than Ruth" or any of that...I start with the LOGIC...that is all.

I find it quite irritating when people claim that all stats are purely subjective...when eventually I'm able to fully explain PCA, I'll let the readers be the judge of how subjective it is...obviously I've had to make choices to set up the model..but they're not based on any agenda with respect to htep layer...they're based on pure logic...

[cubbieinexile]

Memories are faulty and because of the modern technology it has become easier and easier to show that. Not fans but players coaches and other human. There have been numerous interviews or quotes that have circulated for years about players past deeds but when they looked up they find that either parts or all of it is wrong. Currently there is an interview thread in which a poster is putting on past quotes from players, and one of them was asked his fondest memory and it was striking out the side in the 9th inning with Larry Bowa being the final out. The only problem is that according to the stats that never happened.

Then there is the problem of human growth and maturity. A problem that was mentioned briefly in this thread and Bill James talked about. If you are a 20 year old kid standing at the plate for the first time against Bob Feller you are going to think that fastball of his is amazing and that is going to stick with you for the rest of your life. Now fast forward to when your are 35 or 36 and have seen countless pitchers come and go along with their fastballs, would it be as easy to leave the same impression on you as did Bob Feller when you were 20? Not only that but the young coming up in the game are your competition and since you did not grow up idolizing these kids they are not gods either.

[ElHalo]

I believe that looking at baseball purely statistically just isn't logical. Let me explain.

While it may be true that baseball involves a series of finite events that lead inevitably to either outs or bases, and that the way those events occur can be described mathematically... there's so much more to it than that.

Think about the events that go into an at bat. Think about the minute differences in the pitcher's grip that lead to different rotational inertias on the ball. Think about the various levels of skin oil on the fingertips that affect the slide of the pitch along the release point. Think about the humidity in the air and the air pressure that affect the drag on the ball as it travels to the plate. Think about the motion of fast twitch muscle tissue in the batter's arms and legs as he goes into his swinging motion. Think about the imperfections in the surface of the wood of the bat that affect the exact angle of the plane where the bat meets the ball. Think of all the randomness inherent in all of those things, tiny variations that can mean the difference between a solidly hit ball and a lazy pop fly. That's too much randomness for me to believe that you can accurately gain a picture of it all statistically.

Or, do a little experiment. With a fine tipped pen, draw a perfectly straight line down the center of a block of wood. Then take an axe, lift it over your head, and swing down, trying to hit that line. At a guess, I'd say that, no matter how precise you try to make your swinging motion, and looking exactly at the static point you're trying to hit, it would take you thirty tries, at the very least, to get within even five millimeters of the line. Most of the time, you'd probably be an inch or more away from it. Often a lot more. Then think of the difference that swinging five millimeters above or five millimeters below a ball can make. There's so many variables in the reaction of muscle tissue that it's impossible to precisely replicate a swing reliably. Even for the greatest athlete.

To me, there are just far, far too many variables to ever say precisely what the mathematical value of a player is. And so, I generally rely on personal opinion, unless the statistics are so out of line with the perception as to call them into question.

[cubbieinexile]

But heres the thing thing the variables come together to form the stat. If you and I are swinging away at that line we are going to miss and by a lot. But now if we get the 20 best woodcutters in the world we would expect them to be a lot closer and to hit the line a lot more then us. Right? And the stats would show that the woodcutters were better then us. Sure some of their hits might be pure luck but if they/us do it enough times the skill difference between us becomes apparent.

The reason a major leaguer is a major leaguer and we are fans is because they can control the variables better and more often then we can.

[CyNotSoYoung]

Some things might be but some things definitely are not. Anyone who has the opinion that Jose Offerman is a better all-around baseball player than Honus Wagner is, uh,

As jalbright noted - statistics tend to match consensus of opinion. But I'm sure there is some little kid out there somewhere who got Jose's autograph or caught his foul ball or something (and who never heard of Wagner) who thinks he's just the greatest After all, even Joe Schlobotnik had his fan!

[ElHalo]

But heres the thing thing the variables come together to form the stat.

And yes, statistical analysis can of course tell you what will usually happen. However, statistics by definition are useless to tell you what will happen in any particular situation. And if eyewitness accounts say that a player tends to do one thing or another in certain particular situations, then I'll believe them. A lot of players are worse than their statistics. A lot are better. Not usually a huge variance, but it exists. To say otherwise is just denying what's in front of your eyes.