What was Ruth like as a pitcher?

[Snowman]

Originally Posted by Peter_Spaeth

I don't know what you are talking about, but the next 76 at bats easily could have yielded a quite different strike out number for Cobb against Ruth. Not like some inevitable percentage is established by 7 strike outs in just 76 at bats, which easily could have been 5, or 9. I'll let Travis supply the technical explanation but I am trusting my gut here.

He's saying the same thing as you. The statistical distribution of striking out or not is a binomial event (meaning it's either true or false). The p-value gives you the likelihood of getting a 7 Strikeout sample from 76 at bats given his "true" Strikeout rate of 5.6% or whatever it was stated above.

I didn't run the test myself, just going off my gut/experience, but 0.16 passes the smell test.

[Peter_Spaeth]

Originally Posted by Snowman

He's saying the same thing as you. The statistical distribution of striking out or not is a binomial event (meaning it's either true or false). The p-value gives you the likelihood of getting a 7 Strikeout sample from 76 at bats given his "true" Strikeout rate of 5.6% or whatever it was stated above.

I didn't run the test myself, just going off my gut/experience, but 0.16 passes the smell test.

OK that makes sense. I had actually amended my post after rereading his to say I wasn't sure if he was agreeing with me or disagreeing, but yeah I understand it better now thanks to you both.

[EddieP]

Originally Posted by Snowman

He's saying the same thing as you. The statistical distribution of striking out or not is a binomial event (meaning it's either true or false). The p-value gives you the likelihood of getting a 7 Strikeout sample from 76 at bats given his "true" Strikeout rate of 5.6% or whatever it was stated above.

I didn't run the test myself, just going off my gut/experience, but 0.16 passes the smell test.

Based on my line of work that p- value is statistically insignificant.

[packs]

I know Cobb really beat up on Joe Wood and Wood was known to throw exceptionally hard in his time. Maybe Ruth's "tantalizing slow ball" was working for him.

[Touch'EmAll]

Here are consecutive head-to-head duels of Walter Johnson & Babe Ruth. Both were in the American League and played against each other.

August 14, 1915 - Ruth defeats Johnson 4-3
April 17, 1916 - Ruth defeats Johnson 5-1
June 1, 1916 - Ruth defeats Johnson 1-0
August 15, 1916 - Ruth defeats Johnson 1-0 in 13 innings
September 9, 1916 - Ruth defeats Johnson 2-1
September 12, 1916 - Johnson defeats Ruth 4-3

Walter Johnson was in his prime. Babe Ruth was very young. Ruth defeated Johnson 5 times in a row before he lost to The Big Train. Walter Johnson is considered one of the top 3 pitchers of all time, arguably the best.

[Snowman]

Originally Posted by Touch'EmAll

Here are consecutive head-to-head duels of Walter Johnson & Babe Ruth. Both were in the American League and played against each other.

August 14, 1915 - Ruth defeats Johnson 4-3
April 17, 1916 - Ruth defeats Johnson 5-1
June 1, 1916 - Ruth defeats Johnson 1-0
August 15, 1916 - Ruth defeats Johnson 1-0 in 13 innings
September 9, 1916 - Ruth defeats Johnson 2-1
September 12, 1916 - Johnson defeats Ruth 4-3

Walter Johnson was in his prime. Babe Ruth was very young. Ruth defeated Johnson 5 times in a row before he lost to The Big Train. Walter Johnson is considered one of the top 3 pitchers of all time, arguably the best.

This tells us almost nothing about either pitcher's capabilities.

[Snowman]

Originally Posted by EddieP

Based on my line of work that p- value is statistically insignificant.

Based on my line of work, the whole concept of statistical significance is borderline arbitrary nonsense to begin with lol.

[EddieP]

Originally Posted by Snowman

Based on my line of work, the whole concept of statistical significance is borderline arbitrary nonsense to begin with lol.

Hmmmm interesting comment from a “ data scientist” LOL.

[Snowman]

Originally Posted by EddieP

Hmmmm interesting comment from a “ data scientist” LOL.

Let me guess, you think a p-value of 0.049 tells you something about your null hypothesis that a p-value of 0.051 does not? Unless of course you've chosen an alpha of 0.10 (or some other arbitrary value), in which case both are somehow now informative when they previously weren't under your alpha of 0.05?

Frequentist statisticians use a tired framework. If you want to make useful predictions & observations, you should be using Bayesian methodologies.

[nat]

"Like players were putting the ball in play but not hitting it very well"

Yes, I'm pretty sure that's what was happening. But it doesn't really distinguish Ruth from anybody else in the deadball era (Rube Waddell aside).

IIRC, it used to be common for pitchers to throw a knuckleball every once in a while without being dedicated knuckleball pitchers. (Wikipedia confirms, FWIW.) So, it wouldn't be a surprise if Ruth threw the occasional knuckler, even if that wasn't his main pitch.

[Lobo Aullando]

Originally Posted by Snowman

The p-value gives you the likelihood of getting a 7 Strikeout sample from 76 at bats given his "true" Strikeout rate of 5.6% or whatever it was stated above.

One actually needs to take to whole tail, so it's the sum of all the outcome probabilities from 7 Ks all the way up to 76 Ks.

I'll defer on the interpretation, but I'd also note that Ruth struck out 112 left-handed batters in their 733 at-bats against, and Cobb's splits aren't nearly as divergent.

Ruth
Cobb

[Snowman]

Originally Posted by Lobo Aullando

One actually needs to take to whole tail, so it's the sum of all the outcome probabilities from 7 Ks all the way up to 76 Ks.

I'll defer on the interpretation, but I'd also note that Ruth struck out 112 left-handed batters in their 733 at-bats against, and Cobb's splits aren't nearly as divergent.

Ruth
Cobb

Yes, good clarification. It should have read "7 or more".

Also, most stats professors would quibble with my phrasing anyhow. But I was attempting to use layman's terms.