Fixing Inside the Park Home Runs

[hfield007]

Inside the Park Home Runs

as usual sabr has the answers as best will ever be obtained


1901 - 35%

1915 - 25%

1930 - 7.5%

1950 - 3.5%

1979 - 0.9%

1951-present - 0.63%


and cryo, definitive work that is any more helpful than this may be near impossible from what i gathered quick scanning that article. if the guys at sabr dont have it chances are it doesn't exist.


Perhaps use those numbers they came up with and from say 1901-1915 do a gradual even decline for each season to get from 35 to 25.

[CBL-Commish]

Well... it looks like it's going to be some work to get comprehensive, accurate historical lists of ISTP HRs. I just poked around the play index at bb-ref, and I don't see a quick and easy way to mine the data. The data only goes back to WWII, when ISTPers were in rapid decline, and it's not complete. Even up into the 1970s the home run location data has some gaping holes.

One thing we can do, following Ceej's lead, would be to get data on individuals from their HR logs as a starting place. For example, Hugh Duffy hit 19 of his 103 career homers ISTP. Jesse Burkett hit 55 of his 75 ISTP, plus three that bounced over before the ground rule double rule came into effect. Ty Cobb was 46 of 117.

So, my first suggestion would be to have a park effect to cover this. That way we could get it generally right, a lot more right than it is now. And as data becomes available we can enter it for our parks/leagues. This would also allow fictional players to set ISTP HR frequencies that make sense for larger parks.

And I kind of agree with John J. McGraw that we don't have to get this 100% correct right away, just so long as a framework is there to make it right. If Markus sets up a park effect percentage than anyone who doesn't like it can change it, all the way down to 0% of homers are ISTP.

[Cryomaniac]

Originally Posted by hfield007

and cryo, definitive work that is any more helpful than this may be near impossible from what i gathered quick scanning that article. if the guys at sabr dont have it chances are it doesn't exist.

Yeah, that's the problem.

Originally Posted by hfield007

Perhaps use those numbers they came up with and from say 1901-1915 do a gradual even decline for each season to get from 35 to 25.

That makes sense, and is a decline of roughly 0.7% each season.

[CBL-Commish]

Originally Posted by hfield007

Inside the Park Home Runs

as usual sabr has the answers as best will ever be obtained


1901 - 35%

1915 - 25%

1930 - 7.5%

1950 - 3.5%

1979 - 0.9%

1951-present - 0.63%


and cryo, definitive work that is any more helpful than this may be near impossible from what i gathered quick scanning that article. if the guys at sabr dont have it chances are it doesn't exist.


Perhaps use those numbers they came up with and from say 1901-1915 do a gradual even decline for each season to get from 35 to 25.

That's great stuff. And a very nice baseline to start off with.

But I still think we need to do this park-by-park. I would imagine huge differences by stadium in any era. And for my modern setup with a few giant parks that's the only way it's going to work right.

[Cryomaniac]

Originally Posted by CBL-Commish

That's great stuff. And a very nice baseline to start off with.

But I still think we need to do this park-by-park. I would imagine huge differences by stadium in any era. And for my modern setup with a few giant parks that's the only way it's going to work right.

Wouldn't it make more sense to do it globally then modify it with a park factor, which is (presumably) how other stats work?

[CBL-Commish]

We definitely need to continue research here... Just poking around I found that Gavvy Cravath, who retired in 1920, hit only four of his 119 homers ISTP. Babe Ruth hit an ISTPer for career homer number 21, then didn't hit another until career HR 225, and then hit 8 more up through age 34.

Roger Connor hit 17 of 138 ISTP. Dan Brouthers 17 of 106. Sam Thompson 13 of 126. So there were definitely some big sluggers who were well below the era average.

There is some variation by player type and park we should eventually try to capture. But I don't think that should stop us/Markus from implementing a "good enough" solution for now.

[CBL-Commish]

Originally Posted by Cryomaniac

Wouldn't it make more sense to do it globally then modify it with a park factor, which is (presumably) how other stats work?

Yes, I think that's fine. I just want to avoid tying it directly to date or era.

[hfield007]

I don't think we could ever get an accurate representation of it even with just park adjustments because some players in 1901 were more likely to hit more than 35% of their homeruns ITP like speedy line drive hitters like sam crawford (over 50% of his career HR are ITP). unless you tied in speed, gap and power mixed with other things. the big parks weren't the only factor either, much poorer fielding played plenty a role.

in other words, i don't think it's realistic to get this properly implemented all the way through.

[CBL-Commish]

Originally Posted by hfield007

I don't think we could ever get an accurate representation of it even with just park adjustments because some players in 1901 were more likely to hit more than 35% of their homeruns ITP like speedy line drive hitters like sam crawford (over 50% of his career HR are ITP). unless you tied in speed, gap and power mixed with other things. the big parks weren't the only factor either, much poorer fielding played plenty a role.

in other words, i don't think it's realistic to get this properly implemented all the way through.

I'm hoping Markus takes an incremental approach where he starts with park/league effects to get overall percentages right, then maybe in future versions can tie this to player type.

[Cryomaniac]

Originally Posted by CBL-Commish

Yes, I think that's fine. I just want to avoid tying it directly to date or era.

All other stats work like that though.

Originally Posted by CBL-Commish

I'm hoping Markus takes an incremental approach where he starts with park/league effects to get overall percentages right, then maybe in future versions can tie this to player type.

It could be a global percentage, modified by park factor and the player's speed / baserunning / whatever, so a slow player is less likely to hit one than a speedster.

[soundsradio]

If it makes you feel any better, Prince Fielder has two inside the park home runs to his credit. One ricocheted off the metrodome roof.

[CBL-Commish]

Originally Posted by Cryomaniac

It could be a global percentage, modified by park factor and the player's speed / baserunning / whatever, so a slow player is less likely to hit one than a speedster.

That would be fine with me. We don't have comprehensive data, but it seems logical that someone with a combination of gap power/HR power and speed would be the most likely to hit a ISTPer.

And we do have a little data from what I saw yesterday that big slow guys from the 1870-1920 era still didn't hit a lot of ISTPers despite playing in parks with 450, 500 ft. or longer fences. I was a little surprised to see Thompson, Connor, Brouthers, Delehanty, etc only hit 10, 15, 20% of their homers inside the park in that era.

[CBL-Commish]

Dola,

More interesting data... Willie Keeler had 33 career homers, 30 ISTP, 1 bounced. He hit two balls over the fence in almost 10,000 PAs.

Also, I've looked at the top 20 home run hitters through 1920. The vast majority of them hit ISTPers in the 10-20% range, even though the rest of the league was usually far above that mark. A handful of really fast guys with power did better; Sam Crawford was a bit over 50%, Honus was at 46 out of 100, Jake Beckeley at 38 of 87.

But then there are guys like Fred Luderus (who I'll admit to never having heard of before 10 minutes ago) hit 6 of 84 ISTP, but also 62 of 84 in tiny Baker Bowl.

Bad Bill Dahlen was only 10 of 84. Cap Anson 8 of 97. Sherry Mcgee, 22 of 83. Jimmy Ryan 17 of 118.

Fred Pfeffer 9 of 94, but 71 at home including all of his 1884 homers. Lakefront Park in Chicago in 1884 had to have a HR factor of 800 or something. Does OOTP even model that right? That was the park that was < 300 ft almost all around.

Dashing King Kelly only hit 3 of 69 inside the park. Hardy Richardson 11 of 70. Ambidextrous, glove-shunning Jerry Denny only 4 of 74. Laughing Larry Doyle 16 of 74. The last lefty catcher, Jack Clements, only 4 of 76. Home Run Baker only 5 of 96.

In any case... to do this right there's going to have to be a combination of power and speed that determines an individual's ISTP baseline number, and some pretty extreme park effects that are used to split each park's numbers down right. What data we have shows that individual ISTP percentages in the deadball era ranged from roughly 5% to over 90%.

And then we get to the case of bounce home runs, which may have been 3-5% of the total prior to the rules change about 1920ish. Maybe for OOTP15...

[hfield007]

Originally Posted by CBL-Commish

Cap Anson with one of the lowest percentages I found, 8 of 97.

Fun history lesson. Explaining Cap Anson's low ITP numbers.

I would take Cap Anson's ITP numbers with a grain of salt. I refuse to believe any 19th century ITP data is complete or reliable actually.

However, the only reason I may believe Anson's possibly being correct is he played in a home field where in 1884 they made the 200 foot right field fence a legal homerun instead of a ground rule double (the top 4 HR hitters in the NL were White Stockings that year including a quarter of Anson's career homers)

The legal minimum in the 1800's for a fence was 215 feet from home plate and after leaving that park the White Stockings found home at a park with a 216 foot right field fence. Anson hit 2/3 of his career HR at home and only one of those 66 at home without a RF fence of 216 feet or less... which was an ITP

[John J McGraw]

I'm having trouble understanding why getting all these numbers so exact is important or even relevant. I must be missing something and hopefully somebody will post where I'm going wrong.

Whether a home run is inside the park or outside the part makes no real difference. There won't be any more runs scored or even more bases taken. It affects nothing.

So why bother? Well, because it can be added to the pbp to make that more interesting and fun.

Is there anything else? I can't see anything.

In fact, there were so few homers hit each year in the deadball era that there won't be enough to finish with any kind of accurate numbers anyway due to a too small sample size.

Unless there's something I'm missing here, I suggest we make it easy on Markus. Let him set up percentages by decade as close as possible to reality given the lack of real data. Then a bunch of extra lines can be added to the pbp. And it's done.

[hfield007]

The numbers will never be exact (at least not without a time machine and tickets to every major league game where a homerun was hit) and I am hardly worried about an exact indisputable number with this as that is an impossibility. However detail is everything and if we can alter things so they are "more" realistic for historical play I am completely behind it.

[John J McGraw]

Originally Posted by hfield007

The numbers will never be exact (at least not without a time machine and tickets to every major league game where a homerun was hit) and I am hardly worried about an exact indisputable number with this as that is an impossibility. However detail is everything and if we can alter things so they are "more" realistic for historical play I am completely behind it.

That makes sense. But this is really a meaningless detail. Now, I'm not saying make it outlandish. If 1901-1910 averages 28%, don't use 50%. But these stats are totally meaningless except for determining which lines to use in the pbp, right? And since the total sample size is so small, why not just pick a realistic number for each decade? I mean, what's the point of using what might be an exact number and then breaking it down by park? There won't be enough home runs in a year to make that relevant anyway. Why not just make the whole thing easy for Markus to add to the game?

[hfield007]

Originally Posted by John J McGraw

That makes sense. But this is really a meaningless detail. Now, I'm not saying make it outlandish. If 1901-1910 averages 28%, don't use 50%. But these stats are totally meaningless except for determining which lines to use in the pbp, right? And since the total sample size is so small, why not just pick a realistic number for each decade? I mean, what's the point of using what might be an exact number and then breaking it down by park? There won't be enough home runs in a year to make that relevant anyway. Why not just make the whole thing easy for Markus to add to the game?

What you just said supports exactly what I was asking for. I provided the most factual ITP HR numbers via SABR at the top of page 2. I suggested incremental even decreases between the percentages each season starting from the 35% in 1901.

[John J McGraw]

Originally Posted by hfield007

What you just said supports exactly what I was asking for. I provided the most factual ITP HR numbers via SABR at the top of page 2. I suggested incremental even decreases between the percentages each season starting from the 35% in 1901.

And that's a good way to go. I was referring to all the other posts wanting very detailed stats even adjusting for each park. My point was to avoid all that so this can actually get this into the game rather than get bogged down trying to figure out what factor to use in each individual park.

[Ceej]

Today I made an excel spreadsheet covering team totals for the 1890, 1895, 1900 and 1901 seasons and put it up on my Dropbox - you can find it at the bottom of the page here.

I decided to do 1901 to see if I would get a 35% ITP rate in agreement with that SABR article and guess what - not even close! I have it at just over 50% for that year ... I do note that the SABR article in question was originally published way back in 1980 so it's possible that a lot more ITPHRs have been identified since then.... maybe?

I also have 461 home runs for 1901 instead of 454 so I may have double counted some guys who switched teams in midseason, although that alone won't account for the 15% discrepancy. If anyone wants to whip through a couple teams on the file and see if you get the same totals as me, by all means please go ahead.

A few early indications though: I have an overall ITPHR rate of about 19% in 1890 dropping to under 14% in 1895, then climbing to almost 25% in 1900. In addition to Cincy as the article discussed, Exposition Park in Pittsburgh is looking like it was an inside-the-parker haven. Baker Bowl (no surprise) and the Polo Grounds (more of a surprise) are looking below-average for ITPers in this timeframe.