Fixing Inside the Park Home Runs

[Ceej]

File has been updated with 1876-'80, 1884, 1902 and 1905.

Numbers are a bit wild for the 1870s due to few teams, short schedules and franchises coming and going (the league ITPHR rate dropped from 59% in 1877 to 27% in '78, in large part due to Louisville dropping out). But in the 1880s and '90s it's looking like the ITPHR rate was around the 13-18% mark, to put some of the player career figures in post #33 in a bit more perspective.

And then the rate jumps to around 45-50% in the 1900s - this is what I've found for both 1902 and '05 so I don't think it's just me at this point. I plan to do some later decades next to find out more trends over time and get some ITPHR rates for more classic old ballparks.

[CBL-Commish]

Originally Posted by Ceej

File has been updated with 1876-'80, 1884, 1902 and 1905.

Numbers are a bit wild for the 1870s due to few teams, short schedules and franchises coming and going (the league ITPHR rate dropped from 59% in 1877 to 27% in '78, in large part due to Louisville dropping out). But in the 1880s and '90s it's looking like the ITPHR rate was around the 13-18% mark, to put some of the player career figures in post #33 in a bit more perspective.

And then the rate jumps to around 45-50% in the 1900s - this is what I've found for both 1902 and '05 so I don't think it's just me at this point. I plan to do some later decades next to find out more trends over time and get some ITPHR rates for more classic old ballparks.

I'm interested in how some of these rates changed due to park effects. As I previously mentioned, the obvious example is Chicago in 1884 with their tiny, tiny park. That year Chicago and their opponents accounted for 70% of the 8-team National League's home runs, and because of the park almost all of those were in Chicago and over-the-fence.

For example, Silver Flint hit 21 career homers in over a 10-year career. 9 of them were in Lakefront Park in 1884.

Cap Anson had over 10,000 plate appearances spanning 27 years, but 21% of his 97 career homers came at home in 1884 in a few hundred PAs.

Ned Williamson had 64 career homers, 25 of them at home in 1884.

So I'd expect Lakefront Park to have a HR park factor of something like 800 or 1000 (with 100 being average), and an ISTP percentage of 0.5%. The rest of the league would average a HR park effect of 60, with an ISTP percentage of 30%. (All numbers made up for illustration, but still a decent swag.)

[CBL-Commish]

Then you have the interesting case of Buck Freeman. I'd always assumed that his 25-homer season in 1899 was due to a really short porch in Washington. But he had a relatively benign home/road split of 16/9, all but two out of the park. In fact, from 1898-1900 he had 34 homers, 32 over the fence.

After he moved on to Boston later in his career he kept hitting homers in the mid-teens a year, but over 50% of them were ISTP.

[hfield007]

Too many park changes (and outlandish dimensions) and not enough teams in the 1800's to use any raw data at all from the era credibly.

[Ceej]

File has been updated with 1910, 1915, 1920, 1925, 1938 and 1958, as well as a slightly revised 1901 (double counting Doc Newton was the main culprit).

ITPHR rates dropped sharply from 44% in 1910 to 18% at the dawn of the lively ball era in 1920, and were already down to less than 2% by 1938. I have an 0.62% rate for 1958; judging from the SABR article it looks like the rate actually crept back up a little bit in the '60s-'70s, something I might examine next.

Griffith Stadium is definitely confirmed as being a very good ITPHR park, as is Forbes Field and Braves Field prior to the 1928 alteration. Yankee Stadium in its "Death Valley" days was also above average. Not sure why Ebbets Field was good for them in 1915 and '20 but bad for them after that (dimensions didn't change that much in the early '20s)...

[hfield007]

Originally Posted by Ceej

File has been updated with 1910, 1915, 1920, 1925, 1938 and 1958, as well as a slightly revised 1901 (double counting Doc Newton was the main culprit).

ITPHR rates dropped sharply from 44% in 1910 to 18% at the dawn of the lively ball era in 1920, and were already down to less than 2% by 1938. I have an 0.62% rate for 1958; judging from the SABR article it looks like the rate actually crept back up a little bit in the '60s-'70s, something I might examine next.

Griffith Stadium is definitely confirmed as being a very good ITPHR park, as is Forbes Field and Braves Field prior to the 1928 alteration. Yankee Stadium in its "Death Valley" days was also above average. Not sure why Ebbets Field was good for them in 1915 and '20 but bad for them after that (dimensions didn't change that much in the early '20s)...

Somehow I can't help but believe the numbers are incorrect. The article itself references some teams keeping no records of ITP data I just can't imagine it is readily available that no definable work or research exists.

[CBL-Commish]

Originally Posted by Ceej

File has been updated with 1910, 1915, 1920, 1925, 1938 and 1958, as well as a slightly revised 1901 (double counting Doc Newton was the main culprit).

ITPHR rates dropped sharply from 44% in 1910 to 18% at the dawn of the lively ball era in 1920, and were already down to less than 2% by 1938. I have an 0.62% rate for 1958; judging from the SABR article it looks like the rate actually crept back up a little bit in the '60s-'70s, something I might examine next.

Griffith Stadium is definitely confirmed as being a very good ITPHR park, as is Forbes Field and Braves Field prior to the 1928 alteration. Yankee Stadium in its "Death Valley" days was also above average. Not sure why Ebbets Field was good for them in 1915 and '20 but bad for them after that (dimensions didn't change that much in the early '20s)...

I appreciate the work. Hopefully Markus is able to work this in to the upcoming release.

[Ceej]

I did 1966 and 1979 to see whether the HR logs match up with that SABR article and, well, I think it's pretty clear by now that this just isn't gonna be perfect.

I can only find 25 of the 30 ITPs the article says were hit in 1966 - that works out to less than two-tenths of a percent of difference, but it's still kind of grating. There are definitely some anomalies here - the SABR article says Mickey Mantle hit three ITPHRs in 1958, but B-R's HR log only lists two for that season (the SABR article is correct - the "missing" one came on May 9th, it's in the B-R log but simply listed there as a HR). I think based on this, we could surmise that the historical rate is probably a bit higher than what the data I'm pulling out of the B-R logs is saying. The SABR article doesn't seem to be perfect, either - it claims 31 ITPHRs were hit in 1979, but I've found 33 on B-R, plus one by Alfredo Griffin not listed on there.

I did find something else of note while working on 1979. Parks with artifical turf have higher ITP rates than those with grass. This is not really a surprise, what with it being common knowledge that the ball moves faster on turf and gets away from outfielders more easily, as a different SABR article pointed out about the higher rate of triples in turf parks. Nice to see it confirmed, though.

What I guess I'll do from here on in is crunch a couple more old seasons, then do a final post with some general guidelines/recommendations on where to set the rates for all the classic old ballparks through different eras. The numbers aren't going to be perfect but I see enough consistency in the data that I believe we can have numbers in the right, well, ballpark.

[hfield007]

It is why what I stick by what I said, sources are unreliable for ITP homers especially the farther you go back. The data just is not there. This was not something that always got marked in a box score or referenced in news articles.

[CBL-Commish]

Originally Posted by hfield007

It is why what I stick by what I said, sources are unreliable for ITP homers especially the farther you go back. The data just is not there. This was not something that always got marked in a box score or referenced in news articles.

Ok, sure, I get that, and accept it. But we're still far better off incorporating the data we have and incrementally improving the game, than leaving it the way it is. Right now we know that somewhere between 20-50% of homers were ISTP in the deadball era, with the numbers depending on year and ballpark. We also know that in OOTP as it currently stands the numbers for all eras are something like 0.5%.

At the very least, giving us a park effect to set ISTP HR% is going to enable the game to be vastly more accurate in this area than it is today.

[hfield007]

Originally Posted by CBL-Commish

Ok, sure, I get that, and accept it. But we're still far better off incorporating the data we have and incrementally improving the game, than leaving it the way it is. Right now we know that somewhere between 20-50% of homers were ISTP in the deadball era, with the numbers depending on year and ballpark. We also know that in OOTP as it currently stands the numbers for all eras are something like 0.5%.

At the very least, giving us a park effect to set ISTP HR% is going to enable the game to be vastly more accurate in this area than it is today.

I agree a hundred percent with most everything you say. haha. But using hard numbers on years based off of the information gathered is not something I think should be done because given the limited amount of them I don't think it'd properly reflect the odds of hitting one in a given year.

An incremental approach (like I and you have both mentioned) I think is the only reasonable approach. I just was unsure if you were promoting use of hard data each year based off of possibly inaccurate findings.

[Ceej]

File updated - a "couple seasons" turned into doing 1899, 1903, '08, '13, '17, '19, '21, '23, '31 and '48. And this turned up a new peak - 57% in that deadest of deadball seasons, 1908.

I haven't found very much beyond that SABR article to cross-check deadball-era B-R HR log data against, but from what I do have it seems to measure up well. A 1978 SABR article says 34 of 38 homers in Braves Field in '21 were ITP, and that's what I have as well. This book says there were 20 inside-the-park home runs in Yankee Stadium in its first year, but I counted 29. This very extensive write-up on Hilltop Park lists sixteen home runs being hit there in 1903, 1 of them inside the park, which agrees with what I found also. A few home runs are missing from the log however, as I remember seeing some cases where a player has one fewer home run in his log than his official total.

I don't have the time now to post the rest of what I was going to post, so it will follow soon.

[CBL-Commish]

Originally Posted by hfield007

I agree a hundred percent with most everything you say. haha. But using hard numbers on years based off of the information gathered is not something I think should be done because given the limited amount of them I don't think it'd properly reflect the odds of hitting one in a given year.

An incremental approach (like I and you have both mentioned) I think is the only reasonable approach. I just was unsure if you were promoting use of hard data each year based off of possibly inaccurate findings.

Ok, I think we're on the same page. I think that if you're doing a deadball historical sim it would be very good to have a park effect setting where you can just eyeball it and set the ISTP% to 20% or whatever. That's an improvement, and gets you close to real-life results. A lot better than 0.5% ISTPers and an unusual number of fast guys hitting lots of 550-ft homers.

There still will (eventually) need to be a way to tie together ISTP% with gap power/HR power/speed, or you'll still get fairly regular 500-ft homers from 5' 4", 140-pound Willie Keelers.

[Ceej]

file updated with 1893 and 1907 - I'm done now, I swear

Here's a line graph depicting the ITPHR percentages over time according to the data for the seasons I did. The big spike was in 1901 when the American League arrived, and the big temporary drop in 1915 was due to the Federal League having a low rate. The rate declined sharply for good at the beginning of the lively ball era and probably dropped to a little under 1% in the mid-1950s or so. Today in MLB there are usually about 15 inside-the-parkers hit per season, for an even lower rate of 0.3 or 0.4 percent, one would think due to smaller parks and the disappearance of Astroturf. More to come.

[CBL-Commish]

Originally Posted by Ceej

file updated with 1893 and 1907 - I'm done now, I swear

Here's a line graph depicting the ITPHR percentages over time according to the data for the seasons I did. The big spike was in 1901 when the American League arrived, and the big temporary drop in 1915 was due to the Federal League having a low rate. The rate declined sharply for good at the beginning of the lively ball era and probably dropped to a little under 1% in the mid-1950s or so. Today in MLB there are usually about 15 inside-the-parkers hit per season, for an even lower rate of 0.3 or 0.4 percent, one would think due to smaller parks and the disappearance of Astroturf. More to come.

The obvious question here is "what drove the massive spike in ISTP homers in the late 1890s?" The 1920 collapse is pretty easily explained by teams scrambling to move fences in as Babe Ruth's style of play became popularized, along with clean balls and possibly juiced balls. But I don't know of anything that screams out Lots More Inside the Parkers in the 1895-1900 timeframe. I don't think there was massive ballpark turnover in this era, there weren't any huge rules changes between the mound in '93 and the foul-strike rule after the turn of the century. Runs scored were on a gradual downward plane between '94 and 1900, falling off more quickly thereafter.

I really don't know.

Maybe it's related to the drastic, 75% reduction in per-game homers in the 1895-1902 era. Something was causing over-the-fence homers to disappear, and the percentage (if not the total) of ISTPers to shoot up. At first glance it sounds like a bunch of big parks sprung up.

[CBL-Commish]

Of course if I would just read your whole post instead of jumping ahead to the graph, and then read the graph correctly, I'd see the spike was largely due to the American League and a bunch of new ballparks.

But I am curious about the beginning of the spike that seems to have started a couple years earlier.

And it is interesting that it appears the AL had a dramatically different rate of ISTPers than the NL. If the AL was most of the cause of the spike from ~20% to 50% you would have to assume the AL rate was roughly 75%. If the NL stayed constant, that is.

[hfield007]

If I recall correctly, while still around in the early 1900s it wasn't nearly as common as the 1800s to have the outfields roped off (fields weren't permanent, they were wooden and stood for a few years before burning down in most cases) and some have designated spots as ground rule doubles and ground rule triples when a ball rolled far enough to the ropes. I could be way off the more I think about this although the theory is nice. I have no data at all anywhere that i can find of how often ground rules were used for doubles and triples in the 1800s although I know it was not unheard of.


I also considered the fact that fielding improved dramatically over this time but I can't connect how that would effect ITP homeruns

[Cryomaniac]

Originally Posted by hfield007

I also considered the fact that fielding improved dramatically over this time but I can't connect how that would effect ITP homeruns

I assume it would decrease them, since fielders would get to balls hit in the gaps more often.

[hfield007]

Originally Posted by Cryomaniac

I assume it would decrease them, since fielders would get to balls hit in the gaps more often.

exactly why i couldn't connect the two events at all. i can't recall any other rules changes or changes within the game though. Another thought would be if the ball was changed largely during this period which I can't recall.

Other than that, another possibility is the sheer lack of concrete information on ITP from the 1800s.

[Ceej]

So I guess I'll just do these as a rambling series of posts alphabetical by city name, for as many cities as I have time for in a given post.

BALTIMORE
Oriole Park I: 0% (1884)
Oriole Park II: 0% (1890)
Oriole Park III: 20% (1893), 18% (1895), 33% (1899)

These were three different parks all built within blocks of each other, not a ballpark rebuilt twice on the same site. I can't really find anything about the first couple, and 1890 was the only full season that Oriole Park II was even open. According to the scant information on Seamheads the dimensions we know of for Oriole Park III did not change in its eight years of operation - 1899 had a particularly small sample size so an overall rate in the low 20s might be the best place to set it.

Oriole Park IV: 95% (1901), 59% (1902)
This was built on the site of Oriole Park II. It was a weird mix of dimensions being deep in some places (414 feet down the left field line, 480 to right center) and average-looking in others (384 to left center, 330 down the right field line).

Memorial Stadium: 0% (1958), 0% (1966), 0% (1979)
Did not find any ITPHRs here in the years I ran. It was deeper to dead center in its first few seasons, but not remarkably deep by 1950s standards. Probably an average rate would do for this one.

BOSTON
South End Grounds I: was below average for all the 1870s-80s seasons I ran. Rate of about 7% would do it.

South End Grounds II: 6% (1890), 0% (1893)
Center field sure looks huge in that 1893 photo on Wikipedia, and lots of home runs were hit here, just not the inside-the-park kind according to the data on hand. It was short down both lines like the Polo Grounds.

South End Grounds III: 1% (1895), 1% (1899), 5% (1900), 3% (1901), 4% (1902), 0% (1903), 0% (1905), 15% (1907), 0% (1908), 4% (1910), 4% (1913)
So yeah, this park was always well below average for ITPHRs. 3 or 4% would fit for many years.

Huntington Avenue Baseball Grounds: 100% (1901), 85% (1902), 100% (1903), 83% (1905), 72% (1907), 82% (1908), 40% (1910)
A big yard that was excellent for inside-the-park homers in the 1900s. I guess the drop in 1910 must have been because the left field line was shortened from 350 feet down to 305 when the third-base bleachers were extended, according to Seamheads.

Fenway Park: 100% (1913), 33% and 40% (1915), 29% (1917), 23% (1919), 6% (1920), 16% (1921), 12% (1923), 3% (1925), 0% from 1931 forward
Few home runs of any kind seem to have been hit in Fenway in the 1910s, so the numbers are a bit swingy there. It ran about average in the early '20s and then I have no more ITPHRs recorded for the seasons I did, not even in 1931 when it was 468 to center. I would think the only way to ever get one at Fenway in its current configuration would be to hit the ball into that notch just right of center, wouldn't everyone say?

Braves Field: 67% (1915 partial season), 94% (1917), 95% (1919), 94% (1920), 89% (1921), 97% (1923), 93% (1925), 0 and 3% (1921), 0% (1938), 1% (1948)
This was very well known as a great park for ITPHRs (turns out the Braves owner deliberately designed it to be that way). The rate is remarkably consistent in the 90-95% range until the fences were brought way, way in before the 1928 season and it looks like it was probably just average from then on.