Triples+

Yesterday, an HHS reader asked Andy this question:

“Is it possible to determine the greatest triples hitter in history, factoring in era/park, etc.? Wild guess is Willie Wilson.”

Andy passed it on to me, and because the reader is Josh Wilker, author of Cardboard Gods, perhaps my favorite baseball book of the last decade, I thought I’d take a stab at answering the question.  With an assist from @Braves_Paul, I attempted to create Triples+, a metric comparing a player’s seasonal triples total to the league average and adjusting for park factors.

As Josh notes, park and era both weigh heavily into a player’s opportunities to accumulate triples.  Early baseball had huge parks, heavy bats, and fences (if any) so distant that home runs were a rarity, so well-struck balls often left the striker on third.  In fact, in 1901, there were 1.13 triples per game.  In 2013, with strikeouts prevalent and batters swinging for the fences, teams hit an all-time low .32 triples per game.  Clearly, Lefty Davis’s 11 triples in 1901 and Denard Span’s 11 last year were not created equal.

To establish a baseline, I opted to use the same 100-based scale used by OPS+ and ERA+.  This scale is cleaner with rate stats than with cumulative totals, so I started by laying out every qualifying season since 1901 and comparing every player’s triples per game to an average full-time player’s triples per game that season (one-ninth of the average team’s per-game total).  This anointed Curtis Granderson’s 2007 the greatest tripling season of all time, as Grandy tripled 23 times, or .146 times per game, almost seven times what would be expected of a player who received one ninth of his team’s plate appearances.  Thus, Granderson’s season scores a 679, where 100 is average and anything above 100 is better.

Below are the 25 highest seasonal triples totals, adjusted for league average, but not for park effects.

[table id=228 /]

 

It appears Josh was looking for career leaders, but comparing career totals to league averages over each player’s career is beyond my pay grade.  It seems a reasonable substitute could be attained by adding a player’s seasonal Triples+ totals throughout his career in seasons when he qualified for the batting title.  Here are the top 25 by that logic:

[table id=229 /]

 

It’s an admittedly imperfect approach, rewarding triples on a rate basis within each season and accumulation of high-triples seasons.  Then again, the triple is an imperfect outcome, rewarding players who hit the ball not over fence, but far enough from an outfielder that they can run around the bases for a while.  The list above is full of guys who hit a ton of triples, mostly in eras lighter on three-baggers.  And I’m not hesitant to conclude that Ty Cobb was the best tripler ever.

As mentioned above, ballparks have an effect on triples similar to the effect of the game’s evolution.  It’s a lot easier to sprint 270 feet while the ball traverses the vast pastures of Coors Field than to do so in Busch Stadium.  Unfortunately, I’m not sure there’s sufficient data available to make an honest effort at park-adjusting triples.

ESPN actually calculates park factors by event, including triples.  Unfortunately, this data is only available from 2001 onward and it’s far more volatile than I’m comfortable with.  As I understand ESPN’s method, they calculate park factors by comparing triples hit by home and visiting players at the home park in question to those hit by the same players in other parks.  This logic works for frequent events like hits or runs, but an event as rare as a triple produces some wonky results.  Imagine your favorite team only hits 20 triples in a season.  Without an extremely triple-friendly environment, it’s very possible that 16 of those triples could come at home.  Even if visitors accumulate similar numbers of triples at your team’s park and on the road, the home team’s 4:1 home triples to road triples ratio would give the park the appearance of a massive edge in triples.  To wit, Busch Stadium has a .111 park factor for triples (divide a player’s total by .111 to estimate how many he’d hit in a neutral environment) in 2002, but bounces back to .676 in 2003.  Did something happen in St. Louis early in the last decade that made it six times easier to hit a triple there?

Not willing to give up, I chose to regress the park factors a little extra.  Obviously, players play half of their games in parks besides their home, so averaging the home park’s factor with an average park’s factor of 1 is a must (unless one has the time and willingness to look up how many games every player played in every park and create a personalized park factor for everyone).  I chose, somewhat arbitrarily, to give three quarters of the weight to the average park, so as to reduce the impact of the more extreme parks, but not to ignore them altogether.

For parks active throughout the ’90s and into the 2000s, I used the average of each park’s 2001, 2002, and 2003 park factors for every year from 1990 to 2000.  Prior to 1990, I used a park factor of 1 for every player. Of course, I would rather account for actual park factors prior to 1990, but frankly, I don’t know how.  Park factors for triples don’t seem to correlate to park factors for runs (which are available at baseball-reference going back forever) at all.  If anything, there’s a negative correlation, as big parks where homers are scarce tend to yield more triples, but it’s certainly not a direct negative correlation, as parks like Coors love triples and dingers.  If anyone has a suggestion for finding historic park factors for triples, please comment below.

In case you’re curious, the top ten seasons in Triples+, based on real league adjustments and shoddy park adjustments, are as follows:

[table id=230 /]

 

I’d be ashamed to present this list given the inconsistencies in park adjustments, but most of the top seasons are post-2000, which means they’re adjusted consistently.  Comerica Park in 2007 was a triples haven, with a park factor of 1.595, so Granderson drops a few spots, while Reyes sees a Shea Stadium season leap past his top Citi Field Season (and atop the overall leaderboard) thanks to Shea’s .500 park factor in 2008.  Dee Gordon’s 2014, with 8 triples in 72 games, while playing half of his games in a tripler’s hell, would be second on this list if I included it.

Back to Willie Wilson for a moment.  In the 2000s, Kauffman Stadium has typically been kind to would-be triplers, showing park factors above 1 more often than not and peaking at 1.472 in 2009.  Wilson was a flash on the basepaths, capable of turning doubles into triples with some regularity, but he may not have been the best on his team, as Brett liked the three-bagger himself.  That doesn’t mean Wilson doesn’t have a case as the best tripler ever.  My numbers take Cobb.  Traditional numbers take Sam Crawford.  Reyes and Lance Johnson are strong choices as well.  Who you got?

0 0 votes
Article Rating
Subscribe
Notify of
guest

9 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Yippeeyappee
Yippeeyappee
9 years ago

I miss the Cardboard Gods site.

MikeD
MikeD
9 years ago
Reply to  Yippeeyappee

Did it go away when the book came out?

Yippeeyappee
Yippeeyappee
9 years ago
Reply to  MikeD

Inactive for 8 months, though the book came out some time before then. The “community” was really starting to dwindle, as the number of comments per post had dropped right off.

bstar
9 years ago

Even if far from perfect, your attempt at park factors seems pretty well thought out. Great job! One thought I had is to scale the final number, adj triples +, to a number that we can understand. When I see Jose Reyes has the best triples season ever with a score of 674, we don’t know what 674 means. If you multiplied that 674 by a constant so that Reyes’ trips + number matches 36 (the best single-season triples total ever), that would help give the final number more meaning. The downside is you’d be dealing with fractions. Despite that,… Read more »

Voomo Zanzibar
Voomo Zanzibar
9 years ago

(unless one has the time and willingness to look up how many games every player played in every park and create a personalized park factor for everyone)

Does the statistic “AIR” accomplish that?
Or is it only specific to the player’s home park?

Voomo Zanzibar
Voomo Zanzibar
9 years ago
Reply to  Bryan O'Connor

Click on “More Stats”
Third table, “advanced batting”
To the left of Babip

Lawrence Azrin
Lawrence Azrin
9 years ago
Reply to  Bryan O'Connor

#5/B O’C; AIR is similar to OPS+, but it expresses a normalized value for the level of offense compared to the historical average over all MLB history, using .335 OBA and .400 SLG as the average for reference. It is also adjusted for home park. As with OPS+ and ERA+, 100 is exactly league-average; above 100 is above-average environment for hitters, and below 100 is below-average environment to hitters. At the extremes: Hack Wilson in 1930/Wrigley Field has an AIR of 127, Willie Davis in 1968/Dodgers Stadium has an AIR of 68. Nowadays: Mike Trout in 2012/2013 had an AIR… Read more »