1B make more plays (in terms of putouts and assists) than anyone on the field.

They have more chances, but that doesn't mean they make more plays. Catching the throw for a putout isn't "making a play" in the way we normally think of it, though catching a really difficult throw and holding the bag might be. Looked at this way, first basemen do not in fact make more plays than other players; in fact, they make fewer plays now than any position other than catcher (since caught strikeouts are also not "plays" in that sense).

The fact that 2B made fewer DPs in the 19th Century is significant, because there were MANY more baserunners, due to all of the errors.

Yes, it is significant. I don't know the reason for this, other than that defensive ability has obviously improved markedly over time. But while that makes second basemen less important than they are today, I am not sure why that makes them less important than third basemen (other than that is reduces some of their absolute advantage).

Remember that double plays register as two chances for the pivot man. This form of accounting means that lower numbers of double plays will result in lower numbers of chances for those players.

Do the lower numbers of double plays in pre-1920 baseball make shortstops less important defensively, as well? Shortstops are involved in most double plays just as much as second basemen are.

while 2B might have (I'm going from memory) 4 or 5 plays per game compared to 2 or 3 for 3B, the 2 or 3 the 3B makes are MUCH more difficult, considering that there was much more speed in the game during this timeframe, and the throw was much longer and the ball was coming at them faster.

All this still holds true today. Is the argument intended to apply to modern baseball as well?

I should note that there was not "more speed" in the game before 1920. Players today are faster (all athletes are faster) though the fact that they may hit the ball harder would account for some of that difference.

the most convincing evidence is that 2B clearly and easily outhit 3B before 1920 (at about the same level that 3B outhit 2B today).

This is not evidence at all. It may be that the skill set required to play third base was comparatively rarer than that required to play second base, but that doesn't make the defensive importance of the third baseman any higher. It just means that worse hitters will be recruited to play third base.

You have to have a third baseman, just as (in Stengel's immortal phrase) you have to have a catcher. That doesn't mean he will have a lot of defensive value for his team simply because he can actually play the position. It's quite possible, and I think quite likely, that second basemen are still defensively more important (and have a greater impact on winning) than third basemen, even if the defensive skills to play third base are rarer.

Trying to wade into the murky waters of 19th century defense:

First, there were 22% fewer outfield putouts in the 19th century than in the 20th. Why?

1. Because, batters were more inclined to just try to make contact and punch the ball through the infield or BUNT.

2. Offensive strategy involved a lot of stolen base attempts, bunting runners over and trying to take the extra base.

3. 3rd basemen, 1st basemen and CATCHERS all had to be somewhat more mobile than today's players at the same positions because they often had to run in/out to field bunts.

4. Double play ground balls were more rare because of all the stolen base attempts and bunting.

5. In general, more gifted fielders were placed in the infield, with less gifted fielders put into the outfield.

6. Along the infield

<Finally, the most convincing evidence is that 2B clearly and easily outhit 3B before 1920 (at about the

At first reading, I agree with practically all that you wrote, KJOK.

One exception, though. At 1B, I think you're on target for the 1890's & 1900's, but I disagree for the 1880's. I think the typical 1880's 1b-man was both much bigger and somewhat slower than you describe (check out the size of the ABC guys plus Dave Orr, Long John Reilly, Bill Phillips, Mox McQuery, etc.; remember that the average player then was 4" shorter than today, and correspondingly lighter) Good-hands was the prime requirement there.

So who handled the bunts down the 1b-line? I think that might be the pitcher's responsibility; they were 10 feet closer to start with than today, and their typical pitching motion might be less awkward in terms of being in position to field since they were throwing from flat-ground, not a mound.

jimd:

Yes, you're right. I generalized, but there's a good possibility that what I wrote only applies for 1st base from late 1880's-1920 or so, and that in the 1880's possibly and in early decades most certainly, good hands were the #1 priority for a 1st baseman, which means a "big slow slugger" type of player could play the position if he had good hands.

Marc, it (2B hitting better than 3B) proves that they were better defensive players (or the position demanded different, rarer skills -- which means better), because if they weren't, managers could have, and would have put better hitters there.

Joe, it might be a point in favour of your second point, it is NOT a point in favour of your first. You are conflating defensive value with defensive ability, which while it is a mistake with a venerable tradition (the Win Shares defense system relies on conflating these two qualities) is still a mistake.

Look at the defensive spectrum today, it matches perfectly with replacement level for hitters.

Well that's circular reasoning if I've ever seen it. The defensive spectrum, Joe, is essentially defined by the hitting ability of replacement level players. It's not defined by defensive value... in terms of defensive *value* in the modern game, center field appears to be the most important defensive position (don't believe me? Check out any MGL's UZR studies or any other PBP-based measure of defense). It's not the leftmost spot on the defensive spectrum, because hitting talent that can play that position is relatively abundant.

Craig,

Why is factoring defensive ability into defensive value a mistake? I can see that the two things are not identical -- a player can have great ability but not have a chance to use it in ways that benefit the team -- but surely defensive ability contributes significantly to defensive value in most circumstances. I suppose that defensive ability is more useful for comparing players who field the same position, than it is for determining which defensive position is most important, but would you make the case that the ability a defensive position appears to require is irrelevant in determining its defensive value?

but would you make the case that the ability a defensive position appears to require is irrelevant in determining its defensive value?

No, I wouldn't.

But I don't believe that the relative rareness or abundance of the abilities required to man a defensive position are *proof* (remember, my assertion was in response to a claim of "proof") that the players at that position have greater defensive value.

"Actually it does mean that he has a lot of defensive value simply because he can play the position. If not, put Frank Thomas at C or 3B on a 19th Century team and see how many runs you give up. The ability to just be 'there' is the only thing that really gives catchers ANY defensive value, for example, in an era where SB are really low. Yet we still put lowsy hitters there (most of the time) because just being THERE has a lot of value."

Maybe, but to me, the question is this: if Ken Reitz is your 3b and you replace him with Frank Thomas, you are definitely going to give up more runs at 3b, but will you give up more runs at 3b than Thomas will produce at the plate over and above Reitz? I don't know about that. Defensive importance lags hitting importance by a significant margin. Reitz may have defensive value because he can play the position well and Thomas can't, but does that defensive value make up for below-average hitting? Maybe not, in my view.

Plus, you are moving Thomas essentially from no defense (DH) or a worse defensive position to 3b, so it's easy to see why Thomas' defense would be horrendous. But what if you moved a reasonably talented defender from any other position to 3b...I don't know, Dave Winfield or someone like that? Winfield isn't exactly a natural for the 3b position, but he had a good and fairly accurate arm and he was very athletic. Suppose it's 1980 and Winfield produces 91 RC and makes 45 errors at 3b. Are you worse off than having Reitz produce 59 RC and making 8 errors? Yes, if every virtually every error made by Winfield at 3b results in a run for the opposition and if your organization is bereft of guys who can play outfield a little and hit better than Reitz (which would be a poor organization indeed). Otherwise, you probably are not worse off with Winfield there.

So Reitz had some value, but what is that value?

I made this point in another post, but I don't think talent is evenly distributed among the positions over the years. I don't think a manager would ever consider putting Winfield at 3b if he had Reitz, because Winfield doesn't fit the 3b profile. Managers mostly don't use statistical models. They spit, grab their crotch and say "That boy looks like he can play third." In a sense, baseball athletes are typecast. Managers do not necessarily make all the right decisions over time (a la an efficient markets theory). Typecasting positions through the eras is handed down, just like the idea that a sacrifice bunt is a smart play.

"You arent replacing Reit's offense with Thomas' offense, because Thomas was (presumably) already playing 1B. You're replacing Reitz with a second-string 1B. You have to measure the defensive loss of putting Thomas at 3B vs. the offensive gain of replacing Ken Reitz with at best Brian Daubach. Thats not as convincing a case."

I guess it all depends on the players you pick for the example. I agree your example is not as convincing, but using the Winfield example, a team is likely to have a pretty good hitting outfielder somewhere in the system...maybe better than Daubach, and maybe not. But there's always somebody trapped behind somebody else.

"Generally we statgeeks assume that offensive value outwieghs defensive value, but that assumption holds true on the margin. Try to replace Vizquel at SS with Jeremy G. to get Jeremy's bat in the lineup and you will very quickly start leaking runs fast."

Yes, you will leak runs, and in this case, particularly because Giambi isn't very good anywhere and because Vizquel is a pretty decent hitter. You can always substitute the worst defender/best hitter on the team for the best defender/decent hitter and make your argument work. And obviously, SS is a much different example than 3b.

Would be an interesting experiment to run a Strat-O-Matic or "What-if" season for the 1979 Cardinals using Tony Scott instead of Reitz at 3b and letting an aging Lou Brock fill in for Tony Scott in the outfield. You'd lose defense (probably in both places) but gain a bat. I don't think Ken Boyer would consider doing this, because Tony Scott doesn't fit a 3b profile...AND ESPECIALLY because Ken Boyer was a good defensive 3b himself. Seems to me unlikely that he would substitute a good bat at 3b for a good glove, not so much because he understood the mathematical values of runs being created and saved, but because of ego about the 3b position. He would be personally offended by a poor defensive 3b. Of course, I don't know if the Cardinals would have won more games with Scott at 3b. My point is that managers don't always make the most efficient lineup decisions because they are human and there's a lot more to what they do than pure statistical analysis (particularly in the pre-Bill James era).

"You are forgetting all of the extra singles in the hole and doubles down the line Winfield is going to give up. I definitely think it'd be more than a 32 run difference, I think it might be double that. If it wasn't someone would have figured this out and stuck Winfield at 3B."

I agree with your point about singles in the hole and doubles down the line. Whether it would double the difference is an unknown. However, I disagree with the statement that someone would have figured that out and stuck Winfield at 3b. Managers don't manage with the cold efficiency we use to evaluate their decisions in hindsight. Managers and general managers have pretty good instincts, but they've learned the bad behaviors from their predecessors along with the good behaviors, and that's why baseball has a "conventional wisdom". I think there is ample evidence that conventional wisdom is often wrong when statistical methods are employed. Why then would we assume the managers would figure all this stuff out?

This is in response to a debate that KJOK and Chris Cobb have been having on the 1905 Ballot thread concerning Ezra Sutton and Bill Joyce. I put the comment here because I think it has more to do with our next election than our current one.

I have a particular problem with Bill Joyce, and it's going to sound familiar to those who were made uneasy by Rusie's first ballot election.

What is it? Can you name a player with these characteristics?

Managers have tried stretching the D at 3B and it doesn't work too well when they do. The Yanks tried sticking Mattingly there for like a week in the mid-80s. The Giants tried Kingman at 3B for half a season in the early 70s. There are other examples, and they almost always fail, not because the manager won't try it, but because the costs outweigh the gains. They've been experimenting for 130 years, there have probably been over 2500 team-seasons. I'm sure that if it were possible and there were gains to be made someone would have made it work.

The classic example of this stretch working is Sparky Anderson's shift of Pete Rose from outfield to third in 1975 to get George Foster's bat more consistently into the lineup and get somebody else's out. It worked because Rose a) had been a second baseman and b) worked his tail off, in true Pete-Rose fashion, to learn the new position.

I haven't systematically compared Joyce to McGraw, not yet.

I have. McGraw was unquestionably better peak and career than Joyce.

I'm sure that if it were possible and there were gains to be made someone would have made it work.

Coming at this from a different angle: Who were Gil Hodges, Tony Perez, Lee May, Mark McGwire, Jim Thome, Chipper Jones, and Edgar Martinez? They're a bunch of guys who were considered to be third basemen at the moment they moved from the minors to the majors. That's a whole bunch of teams who could have pushed more offense into the lineup had they wanted to by delaying someone's move to a position other than 3B. The list above is far from complete - does Foxx belong on it?

As for the Reds moving Rose to 3B: that meant that Dan Driessen went to the bench and rotted there, because Sparky played his 8 regulars ALL the time.

In 1901, the most common batting position for 2B was 3rd; most common for 3B was 7th. Also, the most common fielding position for 3rd batter was 2B; most common fielding position for 7th batter was 3B. See the entire distribution in "Cumulative Lineup, MLB 1901".