Albright's musings

[jalbright]

The next 10-12 posts are from the thread on using stats versus opinion. 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

[jalbright]

Brothers,

I have a hypothetical scenario for you.

Take your home team, the team you watch the most and root for. Let's say you have a favorite player on said team. You feel you know him fairly well, due to familiarity over time.

And let's say I study this given player by his stats, and have never seen him play. Who do you think know this player better? Me, by studying his page in the record books, or you, who's seen him for many years.

Who is in a better position to truly understand this player's value, you or me?
Think about it.

Bill

If you're a good analyst and objective, your evaluation would be at least as valid as mine. For one thing, the scenario you pose has me emotionally invested in the guy ("favorite player"). Unless I am very careful, that will tend to cloud my judgment. I'll throw a counter-example at you: John Beckwith. He rarely gets the credit I think he deserves precisely because he was hell to get along with. That kind of player has to be Barry Bonds, Rickey Henderson or Ty Cobb great to get his due--and maybe even then he won't. But if he's a great guy everybody loves (think Catfish Hunter) he'll get credit for everything he did well plus some things he didn't even do well, just because he was that kind of guy. Besides, it's only through the stats you can pick up on fine distinctions--for instance, a thirty point difference in batting average is what, about one hit a week? There's no question we should also consider contemporary opinion--it may well pick up on something not well recorded in the stats.

Jim Albright

[jalbright]

Chris,

I would be, because I would never become familiar with a player without also becoming familiar with his statistical record. That would mean I saw him play, and knew about his stats.

Check mate, Chris, I win. Because all those past generations of fans also had access to the stat record. They had both, we have only the stat record.

I win! Sweet!

Bill Burgess

Bill, wile you certainly have a point here, you've captured a pawn, not checkmated. That advantage might enable you to win, it might not. Old time fans didn't look at walks very often--but they're important. They tended to highly overvalue average, which is how DiMaggio could be regarded on a par with Williams when Ted had more power and walked twice as often. Yes, Joe's defense was far better, and he had a worse park to hit in, but those things still don't close the gap. How often did fans of the past even consider park effects, much less have an accurate conception of the size of those effects? If fans of the past had that kind of understanding of the stats, I'd agree with you--but they clearly did not. That's why I had my huge caveat of carefully removing the biases from the data.

Jim Albright

[jalbright]

When we're talking about the opinions of McGraw or Mack or a few guys like them, they certainly deserve a great deal of respect. But most contemporary opinions we get are from less capable baseball men. I was talking more about that mass of opinion rather than the McGraws and Macks. Furthermore, were Mack and McGraw always accurately quoted in context? Of course not. Also, wouldn't you agree that quotes from Mack in the late 1940's or 1950's probably don't carry the same impact as ones made earlier?

I'll concede that if you can put the opinions of two or three guys like Mack and McGraw as opposing what the numbers say, I have to really think about which side is right. I do insist on more than even one such man's opinion simply because no one is always right.

Jim Albright

[jalbright]

I think those of us on the stats side of this discussion often tend to forget that: 1) The data is imperfect--some of it wasn't counted, and even today there are things that are hard to get at in the statistical record and 2) Even the process of evaluating the statistical record requires us to make certain subjective decisions about what to weigh and how much weight to give to certain things. Yes, we can correlate and do all that mathematical modelling, but there's still art in getting the best answers out of the data. It's only when we acknowledge this that we can come up with the very best answers.

Jim Albright

[jalbright]

Properly interpreted stats match the consensus of opinion the vast majority of the time--indeed, that is a true test of the quality of interpretation. When they vary, we should consider which side is off the mark. Every statistical system I have ever seen or used has some scratch your head outcomes. But overall, I find them more reliable than the subjective data we have available. Memory is a very strange beast, as research on eyewitness testimony has shown. When you've got an eyewitness saying the guy was there and his fingerprint/footprint in the victim's blood at the scene--it's a slam dunk. But if the witness disagrees with that kind of evidence, I'm going to look for a reason why he/she likely is making a mistake. If I can't find it, I'll consider that the other evidence could somehow be faulty.

Jim Albright

[jalbright]

For many of us in the debate over stats versus opinion, at least, we're really arguing about approach. Bill Burgess uses stats, but relies on opinions more. I use opinions, but rely on stats more. The key point is we use both. If I have to choose, I will tend to go one way, Bill another. But that's the only difference. If you ignore a whole category of evidence, the greater the possibility is your conclusions will be faulty. There's no similarly "right" answer to what emphasis to give to those categories.

Jim Albright

[jalbright]

The best sabermetric minds are all about finding the truth, wherever it may lie--be it in the stats or elsewehere. There certainly is a group of folks who have worked so long and hard on perfecting their own system that they lose sight of the imperfections in it. At least they're trying to make sense of it all in an organized, thoughtful fashion. I strongly prefer that to seat of the pants evaluations of "emotions, passion and love of the game" aspects of players. Not that those things don't exist, but there is a strong tendency by those I've seen pushing such topics to overvalue those items by factors of at least hundreds.

There will always be some room for argument, no matter the rating system, as we all look for different things. I tend to value career accomplishments more than many others do. Others put greater emphasis on "peak performance"--which itself must be defined. If we somehow try to combine these things, it can only be done in a subjective way. Even if we're both talking about peak performance, if we define it differently say three top years (or four or six) versus top five (our four or six) consecutive years, we can often come up with situations where both sides are right, given their way of defining the term.

The best analysts in sabermetrics are the ones who realize these limitations and spell them out. Then the readers can decide for themselves.

Jim Albright

[jalbright]

Bill James has also been quoted in his book Win Shares as saying "the first thing I did when I checked my results was make sure guys like Babe Ruth and Honus Wagner ended up on top..."

So it's happened more than once...I find it repugnant and I do not do that...ever...there is NO room for favoritism in PCA.

He certainly did that more than he should have. On the other hand, if Ruth and Cobb and Wagner weren't in the top ten or dozen, there was almost certainly a serious flaw in his rating system which needed to be addressed. At that point, he should go looking for such a flaw.

Having tried to perform a similar bit with Japanese players (though with data no more sophisticated or complete than Bill used for his first two Historical Abstracts), I can empathize. I wasn't looking for Oh to come out #1 unless the evidence pointed that way--but if he hadn't been in the top ten, I'd have figured (correctly, I believe) that there was a serious flaw in my system.

I looked at a whole set of questions like that: did I have the guys who subjective raters had put at the very top of their respective positions rated highly among others at the same position? Were most of the eligible guys in their HOF among the best eligible players in my list? I wasn't looking for precise agreement--but if in general, the answers were no, I'd have been digging to figure out where I went wrong. The thing is, as Bill Burgess aptly pointed out in this thread, the fans, historians, managers, front office men and writers who made these ratings had access to the stats and all of them made some use of that data. Furthermore, the stats are a record of what happened, and should contain evidence to support those subjective opinions at least in a general way. Those subjective opinions can easily be wrong in specific cases, but in a more general way they are valid.

Jim Albright

[jalbright]

There's an issue that I don't recall seeing addressed in this thread about opinion evidence. Unless one is willing to do a good deal of research, it's hard to get anything out of them. Until the net made sources much more available, it was even difficult to do. By contrast, there are sources readily available which give us Win Shares, OPS+, Linear Weights, etc for every major leaguer ever. What's more, those statistical evaluations are generally quite good. You get a lot more bang for your buck, so to speak.

Even if you have to input data, as I did for my work on Japan, you can cover a lot more ground in the same amount of time via statistical methods. There's no way I would know anywhere near as much as I do about Japanese ball without heavy reliance on statistical methods. I don't read Japanese, so I'd have had limited access to opinion evidence. I couldn't have easily learned to evaluate the quality of opinions quoted--and a lot of folks using opinion evidence (like sportswriters) just throw quotes around to try and prove their point du jour without ever putting all that opinion evidence into any proper context. You get served slop via opinion evidence often enough, you tend to associate opinion evidence with slop. That's terribly unfortunate for careful, diligent researchers like Bill Burgess.

Lord knows, even today's statistical record in the major leagues doesn't cover everything , and the methods aren't perfect. But they're generally reliable and efficient--that's a big part of why they're so popular.

Jim Albright

[jalbright]

Since I've said I look at subjective data but haven't specified what impresses me (beyond schlarship of the type exhibited by Bill Burgess), I think it's time to put those cards on the table. They don't help Bill Burgess much, because few of these things go back to deadball days and beyond.

I am clearly impressed by strong records in award voting over a period of time and with the BBWAA voting for the HOF. In fact, I have five guys in my queue for the BBF HOF who were selected by the writers but would have been left out by my numbers. I could see the case for all five, and they're in, though not highly placed. Another example I recently came across is that Herman Long finished in a tie for seventh in the voting for pre 1900 candidates in the intial vote for Cooperstown, ahead of numerous HOFers. Multiple selections to post season all-star teams or by the fans to the All-Star game are also impressive. I may not always agree, but I have to give that kind of evidence serious respect.

Jim Albright

[jalbright]

It's interesting news that Cooperstown will be taking up the issue of the Negro Leagues this winter. It's a plus for supporters of those candidates (of whom I am one) in many ways. As I understand it, it has the positives of making room for managers and executives (like Ed Bolden, Cum Posey, J. L. Wilkinson, C.I. Taylor, Vic Harris and Frank Warfield) and having as another focus stars of the era before the leagues were formed (like Louis Santop, Frank Grant, Pete Hill, Home Run Johnson and Cannonball Dick Redding) in addition to the not yet honored stars of the league era (like Cristobal Torriente, Mule Suttles, Alejandro Oms, John Beckwith, Biz Mackey, Jud Wilson and Ray Brown). It's unclear what they will/can do with the abbreviated careers of Chino Smith and Dobie Moore, but I can live with that.

I do have a concern as to how free this well-selected panel will feel to elect a large number of these guys (and maybe some others, especially from the pre-league days) in a one-time vote. I've listed 6 contributors and 12 players (not counting Moore and Chino Smith) I think are deserving. Will they feel free to put in so many in one shot?

From the perspective of my campaign for Japanese candidates, it has plusses and minuses. The precedent of this special committee is a good one for Japanese candidates. The minus is one I won't play up, because it's not right to make those victims of the racism of the color bar wait for Japanese candidates. However, it was a hope of mine that if the campaign for Japanese candidates ever gets traction, we might be able to make common cause with Negro League supporters who were pushing for more of their candidates. The likelihood of that scenario has dropped dramatically with this development.

Jim Albright

[jalbright]

Borrowing from another thread:

I recently borrowed Marshall D. Wright's "The National Association of Base Ball Players, 1857-70," from a local library. Wright's book may represent the only statistical information available from baseball of that era.

Here, I will provided listings from that book for certain star candidates from that era, as well as those already enshrined in the BBF Timeline Hall of Fame.

Some notes about the format I used to give the information:

* The record of a player's team is listed for each year. In 1869 and 1870, the total record and the record against other professional clubs are both given. The city in parentheses refers to the city the club was located in.

* The "Competition" line refers to the area that the team's opponents came from. "NYC Area" takes in New York, Brooklyn, Morrisania (now part of the Bronx), and Northern New Jersey (Hoboken/Jersey City/Newark/etc.).

*From 1857 until 1867, the only information available consists of "runs" and "outs." Anything fewer than 3 outs per game is good. A leadoff hitter might end up with a bit more than 3 outs per game, but Wright's book doesn't give information about batting orders.

* To put the numbers in some context, I quote Wright:



*Finally, after the run, hit, and base totals for each player, I list how good the player was compared to other team members, and note the team's leaders in those categories. Comparing the numbers to NA norms can be misleading because teams did not play standard schedules, and a team's schedule could easily affect the run totals of each player.

*In some cases, I refer to an average and over system. Here's how it works: if Smith scores 33 runs in 10 games, is average would be 3 and 3, or an average of 3 and "3 over." If Smith scored 15 runs in 7 games, he would have an average of 2 runs per game and 1 over.

[jalbright]

For comparison purposes, here's the pre-1871 record of Timeline HOF honoree George Wright:

GEORGE WRIGHT

1864 – Played for Gotham (New York), 3-7-1
Competition: NYC Area
Position: C
Runs: 19 in 8 games. Tied for team lead in total runs with Andrew Gibney.
Outs – 2.37 per game.

1866 (a) – Played for Gotham (New York), 4-4
Competition: NYC Area, Washington
Position: C
Runs: 21 in 5 games. Tied for team lead in run average.
Outs: 1.80 per game.

1866 (b) – Played for Union (Morrisania), 25-3
Competition: NYC Area, Connecticut, Albany, Philadelphia
Position: SS-C
Runs: 42 in 12 games.
Outs: 1.33 per game.

1867 – Played for National (Washington DC), 29-7. Statistics from just 30 games are available.
Competition: East coast from NYC Area to Washington; some teams in the Midwest.
Position: 2B-SS-P
Runs: 182 in 29 games (first on team; George Fletcher had 169).
Outs: 2.21 per game.
Wright’s runs average of 6 runs, 8 over was second among the NA’s established clubs.

1868 – Played for Union (Morrisania NY), 37-6
Competition: East Coast and Midwest
Position: SS-2B
Runs: 195 in 43 games (top on team)
Outs: 2.11 per game.

1869 – Played for Cincinnati (Cincinnati), 57-0, 19-0 against pro teams.
Competition: East and Midwest; five games in San Francisco
Position: SS-P
Runs: 339 in 57 games (first on team; Waterman 293)
Hits: 304 (first on team: Waterman 228. 5.33 per game; nobody else on team had 4 per game.)
Total Bases: 614 (first on team; Sweasy second with 422. 10.77 bases per game; Sweasy had 7.40.)
Outs – 2.03 per game
Wright had 14 IP, and gave up 11.57 runs per 9 innings. No ERA data available.
Wright set NA records for runs average, hits average, and total bases average.

1870 – Played for Cincinnati (Cincinnati), 67-6-1, 27-6-1 vs. pros (best pro team)
Competition: National
Position: SS
Hits: 248 in 58 games.
Total Bases: 411
Wright’s hits average of 4.27 per game and total bases average of 7.08 per game led the NA.

My approach is to compare runs to outs. According to what Marshall Wright has said, an average player will have about 2 runs per game and 3 outs, for a 0.67 ratio. Good players will be at 3 or more runs per game and less than 3 outs, or over a ratio of 1.00. For his first five seasons (through 1869), then, George Wright did phenomenally well, with 756 runs and 299 outs, a ratio of 2.53.

Jim Albright

[jalbright]

Joe Start

Start has been on my Timeline ballot from the beginning. His post-1870 record is well documented. From 1876 until his retirement, he averaged 25 win shares per 162 games, which is pretty good considering he was 33 in 1876. He was the oldest player in major league baseball for the final eight years of his career, and was a productive regular for seven of those eight years.

Start was also considered one of the best players of the 1860s. He was known for hitting the longball; unfortunately, total bases were not recorded as a statistic until 1868.

Data for his pre-1871 career are given below:

1860 – Played for Enterprise (Brooklyn), 2-7
Competition: NYC Area
Position: 3B-1B
Runs: 13 (tied for third on team) in 6 games. (R. Cornwall and Oddie had 16 runs each).
Outs – 2.50 per game.

1861 – Played for Enterprise (Brooklyn), 5-4
Competition: NYC Area
Position: 1B-3B
Runs: 29 in 7 games. (Third on team; Fred Crane and John Chapman each had 30 runs in 10 games.)
Outs – 1.71 per game.
Start’s average of 4 runs, 1 over per game tied him for the NA lead with Campbell of Eckford (Brooklyn).

1862 – Played for Atlantic (Brooklyn), 2-3
Competition: NYC Area
Position: 1B
Runs: 6 in 4 games.
Outs: 2.75 per game.

1863 – Played for Atlantic (Brooklyn), 8-3
Competition: NYC area, Philadelphia, Princeton NJ
Position: 1B-OF-SS
Runs – 23 in 9 games (third on club; Charles Smith had 33 in 11 games, and Dickey Pearce had 30 in 11 games).
Outs – 2.89 per game.

1864 – Played for Atlantic (Brooklyn), 20-0-1
Competition: NYC Area, Woodstock ON, Princeton NJ, Philadelphia, Rochester NY
Position: 1B-3B
Runs – 82 in 18 games. Fifth on team in runs per game.
Outs – 2.61 per game.

1865 – Played for Atlantic (Brooklyn), 18-0
Competition: NYC area, Philadelphia, Washington
Position: 1B
Runs: 82 (top on team; Fred Crane had 71 and Charles Smith 70) in 18 games.
Outs: 2.17 per game
Start led the NA in both runs and runs average this season.

1866 – Played for Atlantic (Brooklyn), 17-3
Competition: NYC Area, Philadelphia, Boston
Position: 1B
Runs: 69 in 16 games (First on team; John Chapman also had 69 runs, but in 18 games).
Outs: 2.31 per game

1867 – Played for Atlantic (Brooklyn), 19-5-1
Competition: NYC Area, Philadelphia, Rochester NY
Position: 1B
Runs: 83 in 19 games, second on team (Fred Crane 88 in 25 games, Pearce 83, Bob Ferguson 82). Only player on team to have a runs average over 4.
Outs: 2.11 per game.

1868 – Played for Atlantic (Brooklyn), 47-7
Competition: East and Midwest
Position: 1B
Runs: 235 in 52 games. First on team in runs and run average.
Hits: 233 (First on team).
Total Bases: 283 (Third on team. Ferguson 312, Chapman 301)
Outs: 2.35 per game.
Among NA teams that kept records of hits, Start finished first in the NA in total hits and hit average (4 hits, 25 over).

1869 – Played for Atlantic (Brooklyn), 40-6-2, 15-6-1 vs. pros (second of 12 pro teams)
Competition: East Coast and Cincinnati
Position: 1B
Runs: 202 in 46 games (First on team; Chapman 197, Pike 193)
Hits: 203 (first on team; Curtis Chapman 197)
Total Bases: 341 (first on team; Pike 325)
Outs: 2.59 per game

1870 – Played for Atlantic (Brooklyn), 41-17, 20-16 vs. pros (fifth best pro team)
Competition: East and Midwest
Position: 1B
Hits: 2.88 per game (best on team; Chapman had 2.58 per game)
Total Bases: 4.41 per game (second on team; Pike 4.58, Chapman 3.62)

Using my runs to outs approach, Start comes in (through 1869) with 824 runs to 468 outs, a 1.77 ratio.

[jalbright]

Dickey Pearce

Pearce is another name on my Timeline ballot. However, the case for his induction rests on his play during the 1860s.

*Pearce was widely regarded as one of the top three baseball players during the 1860s; Start and Harry Wright were in a class with him. However, many of the raves were about Pearce's spectacular fielding - usually calling him the top fielder in the game - and no statistical record of his defense exists.

*Pearce was the captain for the undefeated Atlantic teams of 1864 and 1865. There were basically no managers in those days; the captain controlled the batting order and directed play during the game.

*According to the best research available, Pearce is the person who turned SS into one of the game's key defensive positions. Before Pearce, SS was a position where you would hide a "good-hit, bad-field" player.

*Also, Pearce was the inventor of bunting, and considered one of the best bunters of the day.

*Pearce seems to fill the description of a lead-off hitter; that and his use of the bunt would have increased his outs per game a bit.

Here's the pre-1871 playing record for Dickey Pearce:

1857 – Played for Atlantic (Brooklyn), 7-1-1.
Competition: New York and Brooklyn.
Position: SS
Runs – 28 in eight games. (John Price scored 30 in 8 games; Peter O’Brien scored 29 in 8 games. Mattie O’Brien scored 23 in six games.)
Outs – 2.75 per game.

1858 – Played for Atlantic (Brooklyn), 7-0
Competition – New York, Brooklyn, and New Brunswick NJ.
Position: SS
Runs – 21, good for sixth on the team.
Outs – 3.25 per game.
Pearce is given credit for 8 games, and box scores are not available for all Atlantic games, so Pearce played in several New York-Brooklyn “all-star” games.

1859 – Played for Atlantic (Brooklyn), 11-1
Competition – New York, Brooklyn, and Morrisania.
Position: SS
Runs – 44 (first on team; John Oliver was second with 41. 3.67 runs per game trails only Oliver’s 3.72.)
Outs – 1.92 per game.
Atlantic was responsible for the only loss of Star (8-1) and for two of Eckford’s three losses.

1860 – Played for Atlantic (Brooklyn), 12-2-2
Competition: NYC Area and New Brunswick
Position: SS-C
Runs: 37 (third on team) in 16 games. (Charles Smith had 40, John Price 38). Average of 2.31 was third on team.
Outs – 2.87 per game.

1861 – Played for Atlantic (Brooklyn), 5-2
Competition: NYC Area and New Brunswick NJ
Position: C-SS
Runs: 37 in 10 games. His 3.7 runs per game was second best on the team to R. Seinsoth, who apparently was not in any all-star games.
Outs: 2.70 per game.
The 37 runs lead the NA, and his average of 3 runs, 7 over was third best in the NA in 1861.
Pearce played in at least three “all-star” matches.

1862 – Played for Atlantic (Brooklyn), 2-3
Competition: NYC area
Position: C
Runs: 13 in 5 games. Led team in runs.
Outs – 2.60 per game.

1863 – Played for Atlantic (Brooklyn), 8-3
Competition: NYC area, Philadelphia, Princeton NJ
Position: C
Runs: 30 in 11 games (second on team; Charles Smith had 33, while Joe Start and Fred Crane had 23 each).
Outs – 2.91 per game.

1864 – Played for Atlantic (Brooklyn), 20-0-1
Competition: NYC Area, Philadelphia, Princeton NJ, Rochester, Woodstock ON
Position: C-SS-OF; also team captain
Runs: 94 in 20 games. (Second on team. Charles Smith had 100 runs, and John Chapman had 88.)
Outs – 3.10 per game.
Pearce had the third best runs average in the NA, at 4 runs, 14 over for the season.

1865 – Played for Atlantic (Brooklyn), 18-0
Competition: NYC Area, Philadelphia, Washington
Position: C-SS-2B; also team captain
Runs: 64 in 17 games (fifth on team; Start 82, Crane 71, Charles Smith 70, Chapman 64)
Outs – 3.23 per game.
Atlantic was responsible for two of the three losses by Athletic (Philadelphia).

1866 – Played for Atlantic (Brooklyn), 17-3
Competition: NYC Area, Philadelphia, Boston
Position: SS-OF
Runs: 41 in 12 games (seventh on team; Chapman and Start at 69 runs each, Sid Smith at 50).
Outs – 3.25 per game.

1867 – Played for Atlantic (Brooklyn), 19-5-1
Competition: NYC Area, Philadelphia, Rochester
Position: SS-3B-C-OF
Runs: 83 in 23 games (second on team. Fred Crane had 88, Start 83, Ferguson and Charley Mills 82 each).
Outs: 3.04 per game.

1868 – Played for Atlantic (Brooklyn), 47-7
Competition: East and Midwest
Position: SS-OF
Runs: 191 in 45 games (fourth on team, behind Start 235, Chapman 222, Ferguson 212. Runs average was second on team)
Hits: 185 (fourth on team; Start 233, Chapman 218, Ferguson 194. Second on team in hits per game)
Total Bases: 222 (seventh on team)
Outs: 3.09 per game.

1869 – Played for Atlantic (Brooklyn), 40-6-2, 15-6-1 vs. pros (second of 12 pro teams)
Competition: East Coast and Cincinnati
Position: SS
Runs: 174 in 47 games (fourth on team; Start 202, Chapman 197, Pike 193)
Hits: 175 (third on team; Start 203, Chapman 197)
Total Bases: 236 (fifth on team; Start 341, Pike 325, Chapman 313)
Outs: 3.28 per game.

1870 – Played for Atlantic (Brooklyn), 41-17, 20-16 vs. pros (fifth best pro team)
Competition: East and Midwest
Position: SS
Hits: 2.35 per game (fourth on team; Start 2.88, John Chapman 2.58, Lip Pike 2.48)
Total Bases: 3.00 per game (sixth on team; Pike 4.58, Start 4.41, Chapman 3.62)

Using the runs to outs approach, Pearce has 857 runs to 708 outs through 1869, a 1.21 ratio.

[jalbright]

And now some information for those who played 5 or fewer seasons before 1871.
LIP PIKE

1866 – Played for Athletic (Philadelphia), 23-2
Competition: Philadelphia, NYC Area, New Jersey, Northeastern PA
Position: 3B-OF-2B
Runs: 100 in 16 games (Sixth on team; Dick McBride led with 160 in 25. Runs average was second on team.)
Outs: 3.06 per game.

1867 (a) – Played for Irvington (Irvington NJ), 16-7
Competition: NYC area and Eastern PA
Position: 3B
Runs: 19 in 6 games
Outs: 3.17 per game.

1867 (b) – Played for Mutual (New York), 23-6-1
Competition: East Coast from NYC area to Washington DC
Position: OF-3B-2B-1B
Runs: 82 in 21 games (sixth on team; Waterman 106, Hatfield 100, Hunt 97. Runs average third on team)
Outs: 2.43 per game

1868 – Played for Mutual (New York), 31-10
Competition: East Coast; one match against Cincinnati
Position: OF
Runs: 60 in 27 games. Fifth on team in run average.
Hits: 82 (sixth on team)
Total Bases: 109 (Sixth on team; Swandell 183, Flanly 179, Devyr 174)
Outs: 3.07 per game

1869 – Played for Atlantic (Brooklyn), 40-6-2, 15-6-1 vs. pros (second of 12 pro teams)
Competition: East Coast and Cincinnati
Position: 2B
Runs: 193 in 48 games. (Third on team; Start 202, Chapman 197)
Hits: 175 (third on team; Start 203, Chapman 197)
Total Bases: 325 (second on team; Start 341)
Outs: 2.33 per game

1870 – Played for Atlantic (Brooklyn), 41-17, 20-16 vs. pros (fifth best pro team)
Competition: East and Midwest
Position: 2B
Hits: 2.48 per game (third on team; Start 2.88, Chapman 2.58)
Total Bases: 4.58 per game (first on team; Start 4.41, Chapman 3.62)

DEACON WHITE

White is currently in the timeline HOF.

1868 – Played for Forest City (Cleveland), 11-11-1
Competition: Midwest and Pennsylvania
Position: SS-C
Runs: 73 in 23 games (led team in runs; second in runs average)
Outs: 2.74 per game

1869 – Played for Forest City (Cleveland), 19-6, 1-6 vs. pros
Competition: East and Midwest
Position: C
Runs: 26 in 8 games
Outs: 2.50 per game

1870 – Played for Forest City (Cleveland), 25-16, 9-15 vs. pros (seventh best among pros)
Competition: Midwest and East
Position: C-P
Hits – 3.00 per game; second on team (Ezra Sutton 3.54, Art Allison 2.72)
Total Bases – 5.11 per game (first on team; Sutton 5.05, Allison 4.08)
Pitched 74 innings; gave up 9.61 runs per 9 innings pitched.

Pike won a spot in the Baseball Think Factory Hall of Merit, and White is in the BBF Timeline HOF.

[jalbright]

More info from AG2004 on pre-1871 players:

We also have 1035 runs and 532 outs for Al Reach, for a 1.95 ratio through 1869.

However, the data given above is just for their pre-1871 career; data from the National Association of Professional Base Ball Players (the NABBP's successor, and usually just known as the "NA" today) and the NL are available at baseball-reference.

Reach's last season was in 1875, and he was just hanging on for those last three seasons, so he isn't much of a career guy.

Start lasted until 1886, so he does have the lengthy career. I'd like to see how many people averaged 25 win shares per 162 games after the age of 33, since that's the first season for which we have WS for Start.

Pearce's career started in 1856, but we don't have any records for that year. Baseballlibrary.com quotes the following about him:

The St. Louis Times, June 30, 1868, applauded Pearce, summarizing his contributions : "Pearce has been noted as a superior shortstop for ten years and to-day has no equal in the base ball field. He bats with great judgment and safety..."

Much of the praise for Pearce was for his fielding. We don't have any defensive statistics for the 1860s, but we do have them for the 1870s, and Pearce's career lasted until 1877. In 1874, he led NA shortstops in fielding percentage, was above average in range factor, and was ten years older than any other starting shortstop in the NA. I don't see any evidence there that would lead me to conclude that Pearce wasn't a great fielder during his prime.

There's also an article on Pearce at:
http://www.barnstablepatriot.com/sscape/pearce.html

---

Finally, there are four members of the Hall of Merit at baseballthinkfactory who played at least five seasons before the formation of the NAPBBP in 1871:

*George Wright
*Joe Start
*Dickey Pearce, and
*Lip Pike.

George Wright is already in the Timeline HOF here.

[jalbright]

Lip Pike is quite good in 1866-69, with 1.45 runs per out, and is 4.3 games above average per 162 games by Total Baseball's methods in 1871-75. That covers nine years of his career. Unfortunately, he had 163 games left. He played well in 1876, with 17 Win Shares, but after that, he wasn't much. His career is a little short for someone with a statistical record which needs so much inference and interpretation to my way of thinking

Jim Albright

[jalbright]

Joe Start has an excellent 1.77 runs per out ratio for the 1860's, when 0.67 or so is average and 1.00 is good. In the 1860's, he had 8.41 full seasons. He didn't do too well in the 1871-1875 National Association, averaging 0.7 games above average per/162 games by Total Baseball's methods (Bill James didn't do win shares for the NA). That's the level of a good player, but nothing special. However, he played from 1876 to 1886 and averaged over 25 win shares/162 games there in 9.64 full seasons. That's all-star performance for that last eleven years of his career. A long career with sustained excellence is a recipe for a HOF quality career, even with a slight dip in the NA years in the middle. The Baseball Think Factory guys apparently agree with that thought, since they selected him to their Hall of Merit.

Jim Albright

[jalbright]

Dickey Pearce

He may be HOF worthy, but for reasons I will outline below, I can't go for him. He really didn't do much after 1870, so he's got to make it on the basis of his career before then. I won't argue about his defensive play, either.

However, his runs per out ratio is troublesome. AG2004 suggested in his original posts on pre 1871 players to look how they did relative to their teams because of unbalanced schedules. When I looked at runs per outs ratios for Pearce relative to his teammates, here's what I found:

Year	R/out	Place on team
1857	1.27	5th
1858	0.81	9th
1859	1.91	1st
1860	0.80	5th
1861	1.37	2nd
1862	1.00	2nd
1863	0.94	3rd
1864	1.52	4th
1865	1.16	7th
1866	1.05	6th
1867	1.19	4th
1868	1.37	5th
1869	1.13	7th

Yes, he played for generally fairly high quality teams, though not always (especially when he finished second with a 1.00 ratio). Overall, he's averaging between fourth and fifth on his own team. That just is too shaky a credential for me to back him, even assuming defensive excellence.

Jim Albright

[jalbright]

ELECTED BBF HOF

Deacon White had three seasons before 1871, and in 1868-69, he had a nice 1.24 runs/out ratio in 1.32 full seasons. He followed that up with a 3.1 games per 162 games above average in the NA for 3.97 full seasons using Total Baseball's methods, and then averaged 23.70 win shares per 162 games for 13.18 full seasons after 1876. The latter mark is a solid all-star level. I think that his sustained excellence plus his longevity clearly make him a HOFer. The Baseball Think Factory and Timeline guys have come to that conclusion as well.

Jim Albright

[jalbright]

George Wright ELECTED BBF HOF

He had seven seasons before 1871, and through 1869, he averaged a superb 2.53 runs per out and had 4.60 full seasons. Remember, 0.67 runs per out is about average, and 1.00 is good. He blows those levels away. In the 1871-1875 period, Total Baseball sees him as worth 4.6 games above average per 162 games, which is all-star territory, and he played 4.31 full seasons in those years. He finished with seven years (1876-1882) in the National League, and averaged 25.11 win shares per 162 games in 4.62 full seasons, which is again all-star territory. Thirteen and a half full seasons of performing at an average of all-star level play certainly seems to be an apt description of a Hall of Famer to me. The BBF Timeline and Baseball Think Factory guys agree, having elected him to their Hall of Fame and Hall of Merit, respectively.

If that isn't enough proof for you, from 1871 on, he was in the top 4 in runs scored seven times; the top ten in runs created six times; the top six in slugging percentage four times; the top 10 in OBP three times; and the top ten in average three times. This is from a shortstop, mind you, and it leaves out the first seven or so seasons of his career. He achieved 132 gray ink points in that period, good for 128th best all-time.

Jim Albright

[jalbright]

On one extreme, I looked at the guys already in the BBF HOF

Name......	MLB WS/162	whole MLB yrs	Halls
Cap Anson	27.12......	20.42.......	HTF
Dan Brouthers	34.38......	13.98......	HTF
Jesse Burkett	30.48......	14.58......	HTF
Roger Connor	29.45......	16.50......	HTF
Ed Delahanty	31.34......	13.25......	HTF
Buck Ewing	29.69......	11.03......	HTF
Billy Hamilton	34.31......	11.45......	HTF
Willie Keeler	25.41......	14.84......	HTF

key H=Cooperstown; T=BBF Timeline F=Baseball Think Factory Hall of Merit

Whole MLB seasons are calculated by adding up the percentage of his teams' games a player played in. If a player played in 90% of his team's games in season one, 100% in season two, and 75% in season three, that would come out to 2.65 whole seasons (0.9 + 1.00 + 0.75 = 2.65)

On the other extreme, I looked at numerous guys in one or none of the above-referenced "halls", and the following (among others--unless a guy played before 1875, he had to get 300 or more by multiplying his number of seasons played by his WS/162 and have at least 20 WS/162 to make the cut) were unable to make my cut: Pete Browning, Oyster Burns, Cupid Childs, Abner Dalrymple, Tom Daly, Fred Dunlap, Mike Griffin, Dummy Hoy, Hughie Jennings, Denny Lyons, John McGraw, Cal McVey, Billy Nash, Tip O'Neill, Lip Pike, Jimmy Ryan and Ned Williamson.

Jim Albright

[jalbright]

19th century guys elected to only one or none of the Timeline, BTF Hall of Merit, Cooperstown or BBF HOF

……………….............	MLB……	whole	NA…….	whole	pre1870	pre1870	Other
Name………………......	WS/162	MLB yrs	TPR/162	NA yrs	R/out	wholeyr	Halls
Jack Glasscock….....	24.36	14.94	n/a	n/a	n/a	n/a	F
Herman Long…….....	22.91	13.40	n/a	n/a	n/a	n/a	none
Hardy Richardson....	27.99	11.97	n/a	n/a	n/a	n/a	F
Joe Start………….....	25.17	9.64	1.7	4.92	1.77	8.41	F
Ezra Sutton………....	24.83	11.35	1.7	4.95	n/a	n/a	F
George Van Haltren	28.08	14.19	n/a	n/a	n/a	n/a	T

key H=Cooperstown; T=BBF Timeline F=Baseball Think Factory Hall of Merit

I'd say all six make my cut. Long barely scrapes over, and I'm giving a small break to Van Haltren and Start due to the inexactness of my approach. Start in particular deserves it, though, because his pre 1870 play is so critical to assessing his value, and I've tried to be conservative in that assessment.

The rating system I used for 19th century players is detailed in post #18 of this thread.