36 Triples

[jtur88]

What would it take in the 21st century for a playe to break Owen Wilson's record of 36 triples?

Wilson batted only .300 in 1912, with 175 hits. Only 106 were singles. I am presuming that he played that year with the run-til-you're-out philosophy, since he hit only 19 doubles. Suggesting that he hit 55 gappers, went for a triple every time, but got thrown out at third 19 times. Apparenly, he often didnt even stop at third, as he finished only 3 HRs behind the league leader, But his home run average was about constant through four seasons, so those were probably mostly over the wall.

A generation ago, the triples leader was typically in the mid-teens, and decades would pass without anybody hitting 20. but in recent years, there is usually a player who flirts with or exceeds 20. Relevant to this, though, is that with modern strategy, few players would try to stretch a double into a triple unless there is one out. Risk outweighs the benefit with no outs or two outs. So, unless it ball is still rolling around on the warning track, only about one-third of all doubles would be stretched into a triple. Wilson's Pirates finished in second place, ten games out, but his manager (Fred Clarke) must have tolerated Wilson running him out of a lot of innings. He must have dedicated himself to triples only in that (and the less successful following) season, since the year before, he had 34 doubles, but only 12 triples. He wasn't much of a base-stealer, averaging about ten in an era when dozens of players would steal 30.

Extrapolating from that, one could conclude that a 15-triple player hits most of them with one out, and with disregard for the game situation, might easily reach the 30 mark. I do not know the inning-outs distribution of the triples being hit by recent players with 20---if anybody knows, that would be interesting data. It would also be interesting to see the field configurations of the ballparks that Wilson hit most of his triples in.

Athletes are faster now than in Wilson's time, they wear better designed shoes, and their uniform offers less wind resistance, and they run on a better-maintained base track. (On the other hand, outfield bounces are true on symmetrical fields.) We passed recently through an era when players would steal bases at will, irrespective of strategic benefit. Is it only a matter of time before somebody decides to hit 36 triples? That would require an average of 6 a month, and nobody yet this April has more than 3.

In 1949, Dale Mitchell (Don Larsen's last out) hit 23 triples and only 16 doubles. Was he the last to try to stretch 'em all? In the next half century, only two players (Mays and Willie Wilson) hit 20. Three have done it since 2000.

[Grandstander]

While it fails to explain why others did not capitalize on it to the degree that Wilson did in 1912, nor why Wilson never came anywhere close to the record in any other season, it should be pointed out that the Pirates were playing in Forbes Field and it was 462 feet to centerfield, 376 feet to the rightfield corner, 360 feet to the leftfield corner. There was an awful lot of space for the ball to roll around before it could be policed by an outfielder. 24 of Wilson's 36 triples were hit in Forbes Field, 12 on the road.

Forbes was famous, even back then, for being the best triples park in baseball, while at the same time being the worst park for homeruns.

In addition, triples were easier to come by in those days because several of the venues featured outfields where rather than walls, they had roped off sections behind which were standing spectators. When a ball rolled into the crowd, it was still in play. Until 1926, Forbes Field had bleacher sections in center and leftfield, but the rope situation in right. In 1926, an outfield grandstand was built, shortening the parks enormous fair territory space. In the 17 years before that, the NL triples leader for the season had been a Pirate seven times. (Out of 8 teams) From 1909 through 1925, the Pirates led the NL in triples 8 times, finished second five times and third once. In some of those years they led by immense margins.

Some of the newer parks which have been built, have enhanced triples production, particularly in San Francisco, Houston and Cinncinatti, each featuring odd angles in outfield walls which generate hard to anticipate rebounds. But since the Polo Grounds closed, there aren't any parks remotely close to the conditions which prevailed for Wilson.

[jtur88]

Speaking of doubles and triples, Musial holds one of the most amazing of all records. He led the league in BOTH doubles and triples four times, and led the Majors in both twice. Brock led the majors once. Wagner led his league twice, and Pete Rieser once, but not the majors.

By contrast, Mays never led the league in doubles at all, and was in the top three only twice. Sam Crawford led the league in triples 6 times, but doubles only once.

The record for most doubles and triples combined in a season is held by Joe Medwick, with 64+13=77. Earl Webb was the anti-Owen Wilson, hitting 67 doubles, but stretching it to third only 3 times. (He could have hit 70!)

[jtur88]

GS, your point is well taken about the field configuration, which was obviously a factor that would allow the Pirates to hit a lot of triples. However, if you discount the Wilson anomaly and adjust his totals to his own career typical, the Pirates hit only 28% more triples than the average NL team. In modern times, there is generally a team that hits as much as 50% more triples than the league average, and that team varies from year to year, suggesting it is not a park factor. In '07, the Tigers exceeded the league average by 61%, and in '05 it was the Dodgers by 81%.

Among Wilson's teammates, the 1912 2B/3B ratios, in the same ball park, were Honus Wagner 35/20, Dots Miller 33/12, Max Carey 23/8

[Grandstander]

jtur:

if you discount the Wilson anomaly and adjust his totals to typical, the rest of the Pirates hit only 28% more triples than the average NL team.

28 % is "only?" An average team might hit 75 triples, while the Pirates hit 96...and that's "only?"

[jtur88]

Originally Posted by Grandstander

jtur:


28 % is "only?" An average team might hit 75 triples, while the Pirates hit 96...and that's "only?"

Yes, compared to disparities in recent years of 60-80% on triples. Even for categories like HBP, the team that leads typically has 50% more than the league average. Clearly not park factor.

The point of my question, though, was not how and why he did it, but could somebody else do it under modern playing conditions? When people were leading the league with 12 stolen bases, it might have been asked if Ty Cobb's 96 was impossible under modern conditions. It obviously was not. Could Granderson have hit 13 more, if he had tried in every no-out or 2-out situation where he thought he could make it? In the 50's, when 9 was enough to lead the league, there were plenty of parks with odd configurations.

[Grandstander]

jtur:

In modern times, there is generally a team that hits as much as 50% more triples than the league average, and that team varies from year to year, suggesting it is not a park factor

Aren't percentage disparities greatly enhanced when the overall sampling base is reduced? Last season, the Phillies led the NL with 41 team triples. In 2007 it was the Dodgers with 58. In 1995, 43 triples by the Braves was good enough to lead the league. 1985 it was 58, 1975..54, 1965.. 61, 1955..55, 1945...71..and in 1935 it would have been 68 by the Reds, except that those triple happy Pirates hit 90.

In the 1900's and 1910's, generally, 90 to 100 triples were needed to lead the league and the last place clubs were putting up totals that in many modern years, would have lead the league.

So, with fewer triples being hit overall, it stands to reason that it is easier for one club to have a larger percentage advantage.

[jtur88]

OK, use BBs in 2007. Leader, Boston, 689. MLB average 536. Variance "only" 28%. Would you expect more or less? Do you park adjust it?

Sac Flies: 65, 47, 38%

Strikeouts: 1332, 1073, 24%

Caught stealing: 55, 33, 67%

Errors: 137, 100, 37%

So, a 28% variance does not look like something one would sit up and take notice of, and start searching for ways to park-adjust it, although I have stated that I agree with you that it is almost certainly, to some extent, attributable to the ballpark. Things that can be conspicuously park-adjusted , I would expect a wider variance than that. So I said "only".

Since this is the only point you seem to want to quarrel with, I assume we are largely in agreement on much of the rest of my offering?

[LongtimeBravesFan]

Looking at the all time triples list is interesting. The living player with the most triples is Stan Musial, who is 87 and tied for 19th place with Rabbit Maranville with 177. I'm 40 and the first plaer on the list who played in my lifetime is Roberto Clemnte with 166 in a tie for 27th place. To reach a player I remember playing I have to get to Lou Brock in 63rd place with 141.

The top active player is Johnny Damon in 218th place with 88. Basebal Reference lists Kenny Lofton as active in 106th place with 116, but he is not on a roster at this moment.

To reach the top 1000 all time (actually the top 1030 de to a 33 way tie for 998th place, you'd need only 37 triples. so Wilson almost made the list i just one season. There are three active players in that 33 way tie, so once J.D. Drew, Curtis Granderson and Ken Griffey triple, the ist will be pared down a bit.


Wilson only played nine seasons and had 114 carer triples to put him in a tie for 110th al time

[Triad]

Originally Posted by jtur88

What would it take in the 21st century for a playe to break Owen Wilson's record of 36 triples?

36 triples, while doable in modern times, is not all that likely. I doubt if it will be accomplished over the next 30 years. You mention that players are faster now, but so are the outfielders who chase down the balls. With a ball bouncing off a wall or bounding around in the corner, there's not a lot of time to go 270 feet while the outfielder is retrieving the ball. And there aren't enough quirky bounces during the season for one batter to get him an extra 10-15 chances to get all the way to 3rd.

Point B is that there are fewer triples-type hitters now. More players go for the long ball, which takes away from their double and triple totals.

The easiest way to get a triple is to hit it down the right field line, and then the right fielder generally has to misplay it or have been playing way off the line.

A right-handed batter is not going to hit very many balls down the right field line, so that eliminates about two-thirds of the batting population.

The other less conventional way to hit triples is to hit balls into the gaps, and just be a super-fast runner. And even then, it has to be equally distant from the two outfielders it goes between, otherwise they come up with the ball well before the runner reaches 2nd.

Most triples in a season by active players, with batting direction in parentheses:

Curtis Granderson - 23 (L)
Cristian Guzman - 22 (S)
Jimmy Rollins - 20 (S)
Carl Crawford - 19 (L)
Chone Figgins - 17 (S)
Jose Reyes - 17 (S)
Jose Reyes - 17 (S)
Carl Crawford - 16 (L)
Carl Crawford - 15 (L)

No strictly right-handed batters on the list.

Of the top 20 active career triples leaders, only four are strictly right-handed, and none of those are in the top 10.

Most triples in a season, inactive players, since 1950:

Lance Johnson - 21 (L)
Willie Wilson - 21 (S)
George Brett - 20 (L)
Willie Mays - 20 (R)
Juan Samuel - 19 (R)
Ryne Sandberg - 19 (R)
Garry Templeton - 19 (S)
Willie McGee - 18 (S)
Minnie Minoso - 18 (R)
Garry Templeton - 18 (S)
Tony Fernandez - 17 (S)
Ralph Garr - 17 (L)
Jim Gilliam - 17 (S)
Johnny Callison - 16 (L)
Rod Carew - 16 (L)
Willie Davis - 16 (L)
Paul Molitor - 16 (R)
Jim Rivera - 16 (L)

Only 5 of 18 were strictly right-handed.

I also note the high proportion of switch-hitters on the lists, which only makes sense, because switch-hitters tend to be the speedy types, and they bat left-handed most of the time anyway, so they'll still have lots more opportunities to pull the ball down the right field line.

Active stolen base leaders, with number of 10+ triple seasons, and high in parentheses:

Kenny Lofton - 2 (13)
Barry Bonds - 0 (9)
Juan Pierre - 4 (13)
Omar Vizquel - 1 (10)
Johnny Damon - 3 (11)
Luis Castillo - 1 (10)
Reggie Sanders - 0 (8)
Bobby Abreu - 2 (11)
Carl Crawford - 3 (19)
Ichiro Suzuki - 1 (12)
Edgar Renteria - 0 (4)
Mike Cameron - 0 (9)
Ray Durham - 1 (10)
Alex Rodriguez - 0 (6)
Derek Jeter - 0 (9)
Rafael Furcal - 2 (11)
Carlos Beltran - 2 (12)
Jimmy Rollins - 5 (20)
Gary Sheffield - 0 (5)
Jose Reyes - 3 (17)

Other retired players of note:

Rickey Henderson - 0 (7)
Lou Brock - 6 (14)
Tim Raines - 2 (13)
Vince Coleman - 4 (12)
Joe Morgan - 3 (12)
Willie Wilson - 6 (21)

[jtur88]

I have found the elusive stat, and it is incomprehensibly unaccountable, seemingly defying strategic logic.. Total triples in 2007, MLB, and ratio of doubles to triples:

None out: 335, 10.06
One out: 283, 10.87
Two out: 320, 8.58

Now, somebody please explain: Why would a hitter be LEAST likely to stretch an EBH into a triple with one out? There is clearly a huge strategic advantage to stretching it with one out, as opposed to none out or two out, when there almost nothing to gain by risking a try for third.

Anybody have any thoughts on this?

[nanwynn]

Originally Posted by jtur88

I have found the elusive stat, and it is incomprehensibly unaccountable, seemingly defying strategic logic.. Total triples in 2007, MLB, and ratio of doubles to triples:

None out: 335, 10.06
One out: 283, 10.87
Two out: 320, 8.58

Now, somebody please explain: Why would a hitter be LEAST likely to stretch an EBH into a triple with one out? There is clearly a huge strategic advantage to stretching it with one out, as opposed to none out or two out, when there almost nothing to gain by risking a try for third.

Anybody have any thoughts on this?

I'll take a shot at that.

Ballparks have shrunken considerably in expanse of OF playing areas from the time triples have been cited as being numerous. I picked a personal favorite season of my own, 1941, not too unlike current seasons in terms of scoring and BA numbers. to highlight one contributing factor, batter strike out rates.

1. In 1941, with no DH in play to warp numbers, here's how out distributions looked:

League.......AB.........Hits.........Outs........B A.............K........DP

A.L. .......43125......11492.......31633.... .266 ..... 4425.....1242

N.L. .......42729......11039.......31690.... .258...... 4414.....1152

If we subtract DP's from the out base [assuming duplication of batted ball outs, and knowing that there are some DP's invoving K's] we see that K's
in the AL constituted 14.56% of outs, 14.45% in the NL. That's about one out in seven is a K. Contact was a big deal and batters hated the K. More contact + more OF space = more triples.

2. In 2007, making the same entries as for 1941, we get:

League.......AB..........Hits.........Outs.......B A.............K.......DP

A.L. .......78294......21181........57113... .271...... 14740....2225

N.L. .......89488......23796........65692... .266...... 17449....2429

Doing the same calculations as above for 1941, we get 26.85% K's by out in the AL and 27.58% in the NL, the 2.7% hike in the NL presumably being pitchers batting for pitchers. From one in 7, we've gone to nearly 3 in 10. Smaller OF dimensions + less contact [by half in outs no longer on batted balls] = fewer triples.

The batting climate of all-or-nothing, with high K rates can build doubles and HRs; but 3B become the middle man squeezed out by dimensions, batting dynamics, and financial [and statistical] risk-reward considerations.

3. In addition to park dimension contractions and decreased batting contact, we also face the fact that ballplayers are major investments, with agents and representatives; and millionaire players are a liability if they are injured [increased risk stretching for any extra base, low risk-reward].

4. From the standpoint of odds, or scoring probability, here's how John Thorn & Pete Palmer ["The Hidden Game of Baseball"] presented the 24 base-out situation relative to jtur's question on not stretching for a triple with one out:

Base Situation.........Outs..........Run Probability

none on...................1.............. .249

2B...........................1............. .699

3B...........................1............. .897

If a batter hits a safe double with one out, the statistical odds of his scoring if he stays put are 69.9%. IF he takes a shot at third by stretching for a triple, he MAY up his odds by 28.3% to 89.7%. However, if he loses that gamble, his team faces a reversion back to none on with two out for his failure, or 24.9%.

Our double hitter risks a 64.4% drop in potential in return for a 28.3% possible reward, not a promising return.

Furthermore, in a climate of 27% outs via K, there is little guarantee that being on 3B with one out will score on a batted-ball grounder or sac fly, especially with whiffs being so prevalent.

The risk would have been a better one in former times and ballparks, for a multitude of reasons.

[jtur88]

What you did not account for, though, is the fact that a team with a leadoff double will often give up an out to get him to third, obviously underscoring the advantage of stretching the double to a triple.

I'm assuming that a batter hits a ball that could be stretched to a triple with equal frequency irrespective of the number of outs. If the team strategy is so often to give up a "productive" out to move that runner to third, it would certainly seem that the batter would respect that philosophy and try more frequently (rather than the prevailing less frequently) to get there without necessitating that following out. Granted that there is the risk of losing the baserunner entirely, but the risk remains the same regardless of how many outs. However, the reward increases significantly with one out,, when weighted against the constant risk.

There is, in fact, the oft-repeated adage, that you don't make the first or third out at third base. But hitters violate that, by electing to stretch triples more often, rather than less, in that situation.

[nanwynn]

Originally Posted by jtur88

What you did not account for, though, is the fact that a team with a leadoff double will often give up an out to get him to third, obviously underscoring the advantage of stretching the double to a triple.

I believe you just constructed a 1 + 3 = 5 argument, in that you mixed up your ingredients as if you hadn't done so.

To illustrate what I mean, I'll revert to the 24 Base-Outs Situation chart once again:

1. You now introduce a LEAD-OFF double with the willingness to sacrifice an out to get that runner to third base:

Situation..............Outs..........Run Probability

2B........................0................ 1.068

3B........................1................ .897

none on.................1................ .249

So our doubles hitter has led off the inning; and in the normal course of baseball [as far as statistical probability is concerned] he should score, a situation which inf fact anticipates a rally of sorts.

However, expectation is best converted into reality by sound strategy and execution, so the question arises, "If I sacrifice the runner to third and I now have one out, how are my odds?"

The answer is 89.7% probability of scoring a 16.01% reduction but very nearly even money that the runner will score before the third out is recorded.

You have introduced a new set of parameters [outs; bases] to illustrate how I did not address your first parameters [out; bases].

I'm assuming that a batter hits a ball that could be stretched to a triple with equal frequency irrespective of the number of outs.

That seems logical. The number of outs will not generally affect the free-swinging results of batted balls, if the sacrifice is taken off the table.

If the team strategy is so often to give up a "productive" out to move that runner to third, it would certainly seem that the batter would respect that philosophy and try more frequently (rather than the prevailing less frequently) to get there without necessitating that following out.

That, if you read it carefully, is illogical: the batter-runner, considering the team's "strategy" takes a risk at direct odds with that strategy.

Granted that there is the risk of losing the baserunner entirely, but the risk remains the same regardless of how many outs. However, the reward increases significantly with one out,, when weighted against the constant risk.

I guess all this proves is that you did not read my citation of statistical odds and how they vary dramatically from one base-out situation to another, or, having read it, chose to disregard it.

The numbers presented are not mine: they are the statistically regressed outputs of play-by-play RESULTS of real in-game situations over millions of inputs.

There is, in fact, the oft-repeated adage, that you don't make the first or third out at third base. But hitters violate that, by electing to stretch triples more often, rather than less, in that situation.

You pronounce this general conclusion in a thread you designed to illustrate the exact opposite of your conclusion.

You open with a seasonal record for three-baggers that will probably never be matched, and you further show declines in triples relative to other batted ball effects and then conclude "by electing to stretch triples more often, rather than less, in that situation."

Don't know how any observations could satisfy such apparent illogic. It's like trying to catch time in a bottle.

[jtur88]

Stretching a 2B to a 3B has exactly the same strategic function as stealing third (maybe with trivial variables such as the count on the next batter). However, stretching the hit has the added advantage of the ball already in play in front of the runner when he makes his commitment, so he can stretch or abort after observing play action. Not the case with a steal of third, where the commitment cannot be undone once the break is made, although success or failure depends on pitch type, location and release, and catcher's handling of the ball.

So, if the stretch has the same end result scoring probability as the steal, one would assume that the stretch would occur in the same game situation as the steal of third. Namely, NEVER with 0-out or 2-out. Especially since the runner has a better chance of gauging his success probability when he makes his no-turning-back commitment if he is stretching.