Barry Bonds, 1997 or Babe Ruth, 1927?

The single largest mistake made with trying to compare players from different eras is the insistence in using "Relative to the League" stats. You know the drill. If Ruth is 80% better than the average hitter in the 20s, and if Bonds is 50% better than the average hitter in the 90s, then, goddamn it Ruth MUST be better. But what if the average hitter in the 20s is SO much worse than the average hitter in the 90s?

How can we know if there is no one who played in both eras? Actually, there is.

Part 1 - the pool of players
Let's start the exercise in 1945. And let's look at all those hitters who played in BOTH 1945 and 1946. The first thing we want to do is to limit our pool of players to only those hitters between the ages of 23 to 33. If we allow all the players into our pool, then we might have too many players over the age of 33, which will skew our results, since most "old" players have diminished abilities. By concentrating on a pool of players aged 23-33 we are in effect only counting players who are at their prime.

The next thing we do is to weight the stats by the lower of the PA of the 2 years. So, if you had 20 PA in 1945, and 500 PA in 1946, then we'll prorate the 1946 stats to the 1945 PAs.

Part 2 - the measure of performance
I'm going to use something similar to LW. The plus LW = 1*1B + 1.65*2B + 2.15*3B + 3*HR + .7*BB. (Justification: if you multiply all that by .47, you will get basically the true values of LW. I could have used the run values, but I did not want to get into a debate about runs, and the different effect over time. By using these weights, we are sidestepping that debate. All I want to measure is performance.) The minus LW = AB - H. Take the ratio of plus LW to minus LW, and we get the LW ratio. Simple enough?

Part 3 - Start comparing year x to year x+1
The LW ratio (LWR) of our pool of players in 1945, AL, was .555. That EXACT SAME POOL of players had an LWR of .527 in 1946. This difference is caused by exactly two things: one is the ability of the players to increase, and two is the change in playing conditions (pitchers, parks, weather, strike zones, balls, etc). Since we've already established that these hitters are a stable group in their prime, this implies the difference is COMPLETELY due to the playing conditions. And this means that the PLAYING CONDITIONS caused the players to lose 5% (1 - 527/555) of their hitting stats.

Part 4 - Compare the whole league
The league LWR was .514 in 1945, and .535 in 1946. Again, two things cause the change: playing conditions and talent of the players. In 1946 there was an influx of alot of good players, as you'd expect, both hitters and pitchers. Both these variables caused the hitting stats to go up by 4.1% (535/514 - 1).

Part 5 - Put the two together
Since we know that the playing conditions caused a 5% drop in hitters stats, and since we know that the playing conditions + talent level caused a 4% increase in hitters stats, then we can conclude that the talent level alone was increased by 9.6% (1.041/.950).

Part 6 - Chain it for the whole history
If you continue to follow this methodology, you can get an exact reading of how much the talent level has been increasing or decreasing year-to-year. You simply keep compounding these rates to get an index of talent level. On average there's been a 1% increase in overall talent level from year-to-year. Doesn't sound like much, but when you compound it, it becomes staggering.

In 1927, AL, the talent level was 49% of what it was in 1999. What does this mean? Take all of Ruth hitting stats, and divide them by 2. 60 HRs? Make that 30. 192 hits? Make that 96. 138 BB? No, 69. His LWR was an astounding 1.253. In 1999, it would have been .626. Which just about happens to be the AVERAGE LWR in the AL in 1999.

Based on this methodology then, Babe Ruth would have been an average hitter in 1999. Is it so hard to believe that since he faced such inferior pitching, NO BLACK pitchers, that this is the case? I was a bit surprised by the maginitude.

Shortcomings
Ok, here they are. I did not control for parks. When I took my pool of players, I should have ensured that they played in the same team AND park, so as to remove one variable. I know this is a problem especially when Coors came in, in 1993.

While my pool of players was a stable 23 to 33, I did not control if some years there were more than average number of 23 to 27 players versus 28 to 33. Since the younger group is on their way up, ever so slightly, any imbalance here could skew the results.

AL, 1972/1973. There was a huge gap up in talent level here. This was wholly due to the pitcher not hitting. While the controlled pool of players increased their stats by 3%, the league in its entirety increased by 13%, leaving that 10% increase wholly attributable to the pitchers not hitting and/or the increase in talent level. I would love to re-run my whole system with pitchers' stats removed from the league totals. Anyone who can assist here, by providing pitchers' hitting totals would be appreciated.

As I mentioned, from 1 year to the next, the talent level increased by 1%. Any non-controlled variable could cause a swing to wipe out some of that.

Conclusion
I am not saying, right now, that Babe Ruth would have been an average hitter today, but there's alot of evidence to support such a notion. More work is required.

It looks I'm not the only one going with bold affirmations
1) I'm certain Ruth would not be able to put ruthian numbers today. Pitchers are better today and use better the BB.
2) Ruth was the best player then and Bonds is the best today, so I'm inclined to think they could be equivalent.
3) Ruth would not be an average hitter today because he would have stayed a pitcher instead and a HoF one.

But it's something that came into my mind in the history forum when Aaron and Ruth were compared. In my rating, after period adjustment, Aaron comes in fourth after Ruth, Williams and Gibson (hypothetical though). And Aaron played in the worst time for hitters after the DeadBall Era.
In a way, a doubt is lurking in my mind that maybe Aaron was the best hitter in history. Ruth will forever stay the best in my mind because he could have won 1 or two Cy Youngs (in front of Walter Johnson!!!) if the trophy existed then in addition to a superior hitter, but as a hitter, I don't know.

For me, it is related to the BB. The BB, in my system, worths the same a Bonds or Aaron averages in a year per PA. Of course, you have to isolate the PA's worth of .120 point. But this was another discussion.
__________________
Nouvelles tounes pour récurer les oreilles

No offense, but this is where sabermetrics starts intercepting BS at a very fast rate. It is interesting to speculate on what Ruth would do today, but I believe it doesn't matter.

The argument that athletes have improved over time(decreasing records in every Olympic event), etc. is nice. But I remember someone suggesting once that it's not neccasarily the same for baseball because it combines practically ALL the athletic skills, and specialists don't play a big role.

Of course, I think baseball players are better now than they were in 1920. But the point of baseball is still winning games, and whether Ruth would be great now is not really an issue IMO. At the time he played, he won more games for his team than anyone else ever has. And that makes him the greatest player who ever lived IMO, even if he would hit .230 today.
__________________
"Like all people who reject statistical analysis, they become the captives of the statistics, and wind up making senseless choices on the basis of misunderstood statistics."
-Bill James

I understand the theories. But what I am offering is statistical evidence of the increase in talent level of a baseball player.

Pat: To say that a baseball player is somehow different from a runner, long jumper, a quarterback, etc, simply because he has a larger breadth of skills is something that requires evidence on your part. I'm showing you that there is a mountain of evidence that shows that an average ball player has increased in ability 1% every year. Prove me wrong.

Effel: I understand the Babe would have been a pitcher argument, but that wasn't what I was going after either. I chose the single best hitter in the 1920s to show you how he would hit in the 1990s. The thread titles does say: "Ruth - an AVERAGE hitter today? ". Please, let's not get into semantics here.

By the way, one more shortcoming is that I bundled all the stats into one LWR. Really, I should break it down by offensive event, and give different talent patterns by those events.

In any case, Ruth's 1927 breaks down as follows:
122 H, 18 2B, 5 3B, 38 HR, 88 BB, 442 hitting outs (564 at bats), giving him a LWR of .626, as described previously. You can drops his HR, and increase his hits or whatever you want, as long as his overall LWR comes in at .626.

quote:

---

originally posted by effel
In a way, a doubt is lurking in my mind that maybe Aaron was the best hitter in history.

---

More good stuff from the Tang-master.

Hey Patriot, lighten up. This is more interesting than a debate about whether a 1b is worth .47, or .48, runs.

I assume it should work both ways. If Ruth should get halved, then McGwire should get doubled. So if M McGwire, circa 1998, had played in 1927, he would have had 304 hits for a .597 BAvg, 140 HR for a 1.504 SLG, and 324 BB for a .936 OBA. I find this harder to accept than Ruth being an average hitter today. If all this is true, them pitchers of 1927 were some jive turkeys, man. I'm not sure if this method is valid, but it seems fun to play with.

One thing is that the age range of 23-33 seems too wide, including a few pre and post prime seasons. If players age better today than in the past, that might skew the results. I assume that age range was chosen to maximize the sample size. But it might be worthwhile to repeat the study using ages 25 thru 30 for comparison purposes.

quote:

---

originally posted by Sibelius
I know this is sort of an aside, but Aaron better than Ted Williams?

---

Same problem

But there is something in the formula. If you compare year x to year x+1 and that you repeat it 70 or 80 times, your uncertainties will be greater than what you are trying to measure.

It's funny about McGwire. Everytime I'm trying to picture Ruth, I imagine a McGwire with speed. They both began their career as pitchers.
__________________
Nouvelles tounes pour récurer les oreilles

Thanks, David. I thought I was the only one who found this interesting.

A clarification on how to convert the stats. My basis is always plus events divided by minus events. In that respect, when I say all hitting events should be doubled, I mean that they should be doubled relative to the minus events. So, for Ruth, he had something like 430 plus events, and 340 minus events. He correctly ends up with 215 plus events. However, that brought his overall PA down too low. To bring them up, I had to then multiply all his numbers by 1.30, to put them into their correct form. If you want me to clarify this further, send me an e-mail, and I'll give you a better example.

(The reason I did the ratio of plus to minus is purely for mathematical purposes. If you have a batting average of .400, and you multiply that by 3, you'd get a mathematical impossibility of 1.200. However, if you had a ratio of .400/.600, that gives you a ratio of .67. If you multiply that by 3, that gives you a ratio of 2.0. A ratio of 2.0 translates itself to a .667 batting average.)

Aaron or Williams? If we take 1941 to be the best Ted season, that talent level at that time was 55% of what it was in 1999, AL. Aaron's 1959 NL season faced 74% of the talent level of the 1999, NL. However, Ted's 1957 AL season faced 64% of the talent level of the 1999 AL. At this time, I don't want to make a judgement, because that 1972/73 AL season throws everything off (probably) by 10%. And that might make the difference between Ted and Hank.

Effel, you make a valid point that if you keep compounding, then you introduce a greater error range. And on that basis, you could argue that this methodology is flawed. Like I said, the talent level is increasing by 1% every year, on average. But, if we discover that all my unaccounted for variables really reduced that to 1/2 % every year, well, we all know the effect of compounded interest rates.

David, yes, I used the 23 - 33 group to increase my sample size as much as possible. And there probably are several pre- and post-prime seasons in there. But I did another research whereby that increase in production of a hitter from 23 to 24 is equal to the decrease in production of a hitter from 32 to 33. I was going to re-run the study using only a 5-year age group at some point. I'll let you know the results of that next time.

Richard Cramer ran a study like this about 20 years ago, and people have had lots of problems with it. Was the only issue the fact that he looked at all players, not just players close to their prime? I wouldn't want you to repeat any of the mistakes that Cramer did. But I haven't thought of one yet

[edited by Warren on February 4th, 2001 at 12:38 PM]

Well, Tango, I don't have any evidence, just more knee-jerk thinking.

But the other problem with this as I see it is this. If the talent is better today, why is it better today? Is it because athletic skill in people increases over time? Or is it because people are bigger and stonger now than they used to be?

Or is it the Stephen Gould theory, that baseball players are evolving. People don't have tails anymore(well, I guess nobody knows if they ever did), in the same way that pitchers can't hit anymore? Did pitchers stop being able to hit because it's not critical and they spend their time working on pitching?

The other thing I would suggest is this. Mike Schmidt was 6'2, 195 lbs. Gehrig was was 6'0, 200. So even if Schimdt played in tougher conditions, I can't see any reason why Gehrig couldn't adapt to those conditions and hit, not as well as he did in the thirties, but match what Schmidt did certainly. Whatever advantage or disadvantage Mike had, Lou would have if he too was born in 1949 too, IMO.
__________________
"Like all people who reject statistical analysis, they become the captives of the statistics, and wind up making senseless choices on the basis of misunderstood statistics."
-Bill James

David, I re-ran with a pool of players aged 24 to 28 (which I can prove in ONE respect as being prime; more on this later). What happens here is that this shows that the talent level in the NL has been VERY STABLE from 1919 onwards (with a couple of bumps along the way).

Warren, this brings me up to your point. Yes, Richard Cramer's methodology was correct, but his pool of players was wrong. He considered ALL players, and there are lots more players post-prime, than pre-prime. This has the effect of showing that while your controlled player might be losing 1% of his talent, the league in all might be stable, thereby producing the effect that the talent level of the whole league INCREASED by 1% (just examples here).

David/Warren: so, it becomes critical, ABSOLUTELY CRITICAL, that we determine what is the proper prime years / aging curves. Even worse, is that home run hitters age differently than speed demons. You can even make the case that if the 80s was OVERRUN with these speed demons, that hitters in the 80s would lose their abilities faster than hitters today, simply because they are different types of players.

What Voros has done I believe is to determine what is the aging patterns for each offensive event. By using these "adjustments", you can correct for all this. However, you now have to make a leap of faith that these adjustments are constant over time. That is, a power hitter aged 24 in 1946 will increase in ability to the same extent of a 24 year old in 1996. Some people may not believe that to be a hurdle, and others do.

Pat: As for Gehrig/Schmidt, it MAY be that with conditioning, Schmidt is slightly stronger, faster, more well-rested, etc. (Just using these 2 guys as examples, but you can choose anyone.) My methodology does not show what would have happened if Gehrig were born today, how would he hit. It shows what would happen if you plucked Gehrig out of the 20s, and plucked him into the 80s,and give him the same playing conditions as Schmidt, how would he have done.

In hockey, if anyone follows it, players are not only bigger, but MUCH bigger than even 30 years ago. They are not only faster, but MUCH faster. They may or may not be smarter. They may or may not have better practice regimen. They may or may not have more talent. But I am sure if I were to undertake a similar study, it would show that outside of the elite players (Bobby Orr, Guy Lafleur, etc), that they would be 2nd line players today. And if we go back to hockey in the 30s? Well, i don't even think it would be close.

Baseball has more "precise" type of skills, and a larger breadth of skills. It may be that the talent level has been pretty stable for the last 50 years. I do not know. But I won't say one way or the other, without providing evidence for it, as well as all my doubts.

AL/NL: one interesting thing that I show is that the AL has been a much inferior league to the NL up until about 30 years ago, whereby they have been even. Again, I haven't studied the issue, but this looks like one of the by-products of my study. We need to look at this as well.

Based on Tango's clarification, I recomputed what McGwire's 1998 stats would have been in 1927:

207 H
95 HR
221 BB
.460 BAvg
.638 OBA
1.158 SLG
185 runs scored (est.)
210 RBI (est.)

If you don't believe that 70HR today can become 98HR in 1927, well, what would happen TODAY if you take out all the black and latin pitchers, and replace them with tripleAA and doubleAA white pitchers? No Pedro. No Mo Rivera. No Urbina. Take them all out. All of them. And replace them with 12th, 13th, 14th and 15th pitchers.

And what would happen TODAY if baseball pitchers were not instant millionaires, but rather the poor slobs they were in the old days? Well, you'd have a whole slew of great baseball players who decide to NOT play baseball, and pursue instead their engineering degrees, computer science degrees, and law degrees. Take those guys out, and now you've got the team's 16th and 17th pitchers making the team.

Now, consider, how would Mark McGwire do against that kind of pitching? If someone has the data, here's one way to prove it. Break down the pitchers faced by Mark and 19 other great hitters into 3 equal groups: the great pitchers, the average, and the bums (and take out all the black and latin pitchers). Then remove 10% of all the stats, because the pitchers may have pursued regular careers instead. Now, all of the "removed IP" must be added back in based ONLY on the stats of the bum pitchers. What do you get? Anyone, anyone? Bueller? Anyone?

quote:

---

originally posted by tangotiger
If you don't believe that 70HR today can become 98HR in 1927, well, what would happen TODAY if you take out all the black and latin pitchers, and replace them with tripleAA and doubleAA white pitchers? No Pedro. No Mo Rivera. No Urbina. Take them all out. All of them. And replace them with 12th, 13th, 14th and 15th pitchers.

And what would happen TODAY if baseball pitchers were not instant millionaires, but rather the poor slobs they were in the old days? Well, you'd have a whole slew of great baseball players who decide to NOT play baseball, and pursue instead their engineering degrees, computer science degrees, and law degrees. Take those guys out, and now you've got the team's 16th and 17th pitchers making the team.

Now, consider, how would Mark McGwire do against that kind of pitching? If someone has the data, here's one way to prove it. Break down the pitchers faced by Mark and 19 other great hitters into 3 equal groups: the great pitchers, the average, and the bums (and take out all the black and latin pitchers). Then remove 10% of all the stats, because the pitchers may have pursued regular careers instead. Now, all of the "removed IP" must be added back in based ONLY on the stats of the bum pitchers. What do you get? Anyone, anyone? Bueller? Anyone?

---

Send Barry Bonds back to 1876, and what would you get? There are probably some upper limits to performance; this method might be better used to evaluate average players than superstars.

I have no doubt that the average player of today is better--much better--than the average player of the past. But IF the greatest player ever DID play some time in the past, there is no way this method would find him.

quote:

---

originally posted by Andrew
Send Barry Bonds back to 1876, and what would you get? There are probably some upper limits to performance; this method might be better used to evaluate average players than superstars.

I have no doubt that the average player of today is better--much better--than the average player of the past. But IF the greatest player ever DID play some time in the past, there is no way this method would find him.

---

I think conditioning and training does have a lot to do with this. I think if you magically teleported the 1927 Yankees into 2001, they wouldn't necessarily be the best team. However, if you allowed them to grow up in this society, they'd have the same advantages as today's players.

I guess it depends on your definition - I think in one sense Barry Bonds is the greatest player in history, but have him grow up in the 20s and 30s and I'm not sure he's better than Babe Ruth.

I'm sure integration and the influx of foreign players has also had a very large impact. Imagine an all-star team consisting of only American-born white players today.

I think the methodology is interesting, but I just cannot bring myself to believe in the conclusion. I would like to see some more studying because my gut could be wrong.

Here is my main concern with the analysis: You are taking the average production and of players and then applying the results to outliers. I think a more formal conclusion would be that an average player in 1999 is worth x% more than an average player in 1920. It is probably true that the average player has increased, but I do not see how this applys to the people at the extreme.

Many analysts state that the distribution of players is like the upper half of the normal curve meaning that there are many more .270 hitters than .300 hitters, but noone will hit .210 and be allowed to stay in the league. Your grouping of players will take in so many of the average players that I cannot see the application to the extremes.

An inflation rate will rise from year to year and this might hurt the average family, but it will have no impact upon Bill Gates...he might lose some money, but he will still be rich. In the same vein, the average 1920's players might be worse off in today's game, but I just cannot see Ruth having all his stats cut in half.

I commend the time and effort on the research and there are worse things in life than having me disagree with you. Keep us posted if anything shines through in further studies.

quote:

---

originally posted by tangotiger
If you don't believe that 70HR today can become 98HR in 1927, well, what would happen TODAY if you take out all the black and latin pitchers, and replace them with tripleAA and doubleAA white pitchers?

---

(Ignore. Testing posts.)

post 1 of 3

First, let me respond to those with concerns.

16 teams instead of 30? No. There are 30 teams because there is demand for it. As well, since the population has increased dramatically from 70 years ago, I would guess that the per capita rate is probably constant.

goody14/andrew: yes, you are absolutely correct, and I'll add that to my list of shortcomings. As I've said countless times in other threads, just because you have an average does not mean that that average can be distributed to everyone. If you bring the fences down to 260 feet, will McGwire hit alot more homers? I don't know about alot, but he'll hit more homers. What about Rey Ordonez? Well, all of a sudden, all his easy fly balls now has a better chance of going out. This is just to say that just because you can double the average players stats, doesn't mean that ALL players get their stats doubled.

post 2 of 3

warren: correct, we have 2 separate definitions. My definition was to "magically transport" them, rather than having them "grow up" in the particular time. Chances are Ruth would not have been a drunk, and he would have hit better possibly. Maybe he had injuries that kept him from performing even better, injuries that would be treated today. The important thing is to define our basis, and then study the effects.

rock: yes, absolutely, McGwire could have been finished by 25 in the 1920s, if his injuries could not have been tended to properly. As well, I never said there are no good white pitchers. I said that if you take out a SIGNIFICANT pool of pitchers, you have to replace them with someone. And those someones will all be minor league pitchers.

post 3 of 4

Now here's further fuel to the fire on removing black/latin pitchers. I took ALL the pitchers in 1999, NL, with at least 50IP, and split them up into 5 classes of pitchers, based on ERA. Here are their HR/game rates (best pitchers to worst):
0.82
1.03
1.03
1.22
1.46
Is it fair to say that if we could project what a minor league pitcher would do, we'd get the following for the next 3 classes of pitchers?:
1.66
1.86
2.06
If we assume that each class of pitchers represents 12.5%, instead of 20% of current pitchers, this would give us the "white non-educated" pitchers.
The average HR rate for these white pitchers would be 1.40, instead of the NL, 1999 rate of ALL good pitchers (regardless of race), of 1.11.

quote:

---

originally posted by tangotiger
Barry Bonds, 1997 or Babe Ruth, 1927?

The single largest mistake made with trying to compare players from different eras is the insistence in using "Relative to the League" stats. You know the drill. If Ruth is 80% better than the average hitter in the 20s, and if Bonds is 50% better than the average hitter in the 90s, then, goddamn it Ruth MUST be better. But what if the average hitter in the 20s is SO much worse than the average hitter in the 90s?
Are you asking or are you stating that the hitters of the the 1920s are SO much worse than the hitters of the 1990s. If you are stating that where is your proof. Don't go off on a tangent now and tell me about inferior pitching in the 1920s because we could argue all day about pitching then and now. While on the subject of conditions then and now, consider some of the advantages todays player have over those of the 1920s, expansion, a suspect ball, much smaller parks, a lower mound and a midget league strike zone.You can debate he expansion and the ball issue, but not the smaller parks, lower mound and lower strike zone.So how do you prove hitters of the 1920s are SO much worse than hitters of the 1990s. The only thing hitters of the 1990s do better than hitters of the 1920s is hit HRs. League doubles, triples and most important the league batting average of the 1920s is higher than that of the 1990s.





In 1927, AL, the talent level was 49% of what it was in 1999. What does this mean? Take all of Ruth hitting stats, and divide them by 2. 60 HRs? Make that 30. 192 hits? Make that 96. 138 BB? No, 69. His LWR was an astounding 1.253. In 1999, it would have been .626. Which just about happens to be the AVERAGE LWR in the AL in 1999.

How do you prove the talent level of 1927 was 49% of what it was in 1999, with numbers and math, no way to prove that statement. Now your getting funny, cut Ruth's totals in half, so you have so many hitters hitting 40+ HRs, 50 and 60+ HRs and you come up with Ruth hitting around 30.So the man who hit a HR every 11.76 at bats and still managed a .342 life time average would just be an average hitter today.
Based on this methodology then, Babe Ruth would have been an average hitter in 1999. Is it so hard to believe that since he faced such inferior pitching, NO BLACK pitchers, that this is the case? I was a bit surprised by the maginitude.
Prove beyond a doubt that pitching a was so inferior in the 1920s, it can't be done. The problem here is you base most of your opinions with too, too much math.

---

post 4 of 4 (being prevented from posting in 1 shot)


We can conclude that the AVERAGE hitter TODAY, if he faced only those white pitchers who chose a baseball career would have gotten 25% more HR TODAY. Whether this would translate itself to the great HR hitters as well, I do not know.

rock: let's see. I present a systematic and logical approach of what would happen if you remove black/latin pitchers, and you respond by saying that my kind of thinking is what makes people from taking us seriously. Rather than turning these boards into a "he-said she-said yahoo boards", why not take the time and effort required to analyze the situation, and present your systematic and logical findings?

I invite everyone here to undertake some analysis of any kind to either dispute any of my findings here, or add to them.

admin options:
>open / close topic
>move thread
>delete topic
>edit topic