Play Punch out - Put the Stats In Your Favor

Originally Posted by BaseballInstructor50:

Check out how the statistics change from pitch to pitch when you throw a strike vs ball.  Learn which pitches are the most important and execute in those situations.  Have fun playing Punch Out.

 

I commend you and anyone else who attempts to get people to try to look more into the game by having them look at the numbers. The only question I have is, where did the numbers that appear once a choice is made come from? Are they ML, MiL, College, HS, or some lower level numbers?

 

The reason I ask is, from level to level, the numbers might vary tremendously.

I agree that the numbers themselves will vary from level to level. However I think the overall significance of each pitch will remain consistent respectively.  These numbers were pulled from mlb statistics.  We have to assume that pitchers, hitters, and umpires are all evenly matched well enough to prove the numbers to be mostly true from level to level.

Pitchers do not record outs on 73.9% of the 0-1 pitches they throw. Batting average at a specific count is an irrelevant statistic.

 

Batting average is very relevant in a specific count.  I'd much rather throw a pitch to a .300 hitter with an 0-2 count than a 2-0 count.  If you think batting average isn't affected by the count then you're choosing to ignore the statistics that are there for a reason.

BaseballInstructor50- In your link, you state that a player's batting average on a 0-1 count is .261. This means that on 0-1 pitches, an out is recorded 73.9% of the time. That is not true, and therefore the statistic you cited is incorrect.

 

I never said that statistical splits per count are not impactful. I said that batting average in a specific count is irrelevant. There are other stats for the analysis of batter splits that are far more accurate in illustrating the true measures of success in specific instances.

 

 

Which is why you want batting average (if you want to use batting average) through X count and not at.

 

The key point being "if you want to use batting average." I don't, ever, want to use batting average. But that's a topic for another day.

 

The statistics simply show that hitters have a .261 avg swinging at first pitch. Your assumptions are incorrect that a pitcher records an out on first pitch 73.9% of the time. There are errors that come into play as well as dropped strike three plays.  Once again, the statistics are there for a reason and the punch out game is to show the importance of throwing strikes versus balls in every count

BaseballInstructor50- Batting average on a specific count is irrelevant and has no statistical predictive value. 

 

Batting average in a specific count only factors pitches that were put into play. It does not factor in balls, called strikes, or foul balls. The predictive value of such a statistic (which is really BABIP, not BA) becomes inaccurate due to several factors, such as the volatility of the defense, how well-hit the ball is, the volatility of the pitcher, the volatility of the hitter, and luck. Saying "hitters hit .261 when they put a 0-1 pitch in play" predicts what, exactly?

 

There is statistical significance to batting splits in specific counts, as I mentioned before. If you'd like to cite batting average (a practice I try to stay away from usually), batting average through a specific count would be a much better model to provide statistical predictive accuracy. 

 

Here is a link to an analysis on the topic. The data is from 2006 and therefore is a bit outdated, but the formulaic process doesn't change. The most predictive statistic within this data set is wOBA. http://www.insidethebook.com/e...ts/hitting_by_count/

 

Here is a link to the count-splits on Mike Trout's 2014 Player Page on Baseball-Reference. The statistical sets of value are those that are "After" a specific count. While I prefer Fangraph's wRC+ as the most accurate normalized offensive statistic, tOPS+ is the best on this page: http://www.baseball-reference....r=2014&t=b#count

 

Once again, BaseballInstructor50, I am stating that batting average on a specific count has no predictive value and is therefore a meaningless statistic. Hitting success most certainly differs from count-to-count and I wholeheartedly endorse the concept of teaching this using statistics. However, using batting average on a specific count simply does not illustrate the point, at all.

 

The idea that BA is a worthless stat isn’t something new. Here’s 2 articles written by FC Lane, one in 1916 and the other in 1917. If it was obvious to one of the earliest Sabermatricians, why is it so hard to see now.

 

https://courses.edx.org/c4x/BU...ges_1916_FC_Lane.pdf

https://courses.edx.org/c4x/BU...e_Reform_FC_Lane.pdf

 

Stats- I don't think batting average is "worthless" per se. Stats like Pitcher Wins and Saves are worthless. At least batting average is something that tangibly occurs on the field.

 

However, as noted a century ago (thanks for the links) and continued to be proven since, there are far better statistical measures for offensive performance. I can't answer your question as to why some people don't adopt these measures, I can only continue to understand that statistical analysis of baseball is better because of them. 

 

I will say, as I've said here on this site many times, that I would love to engage in conversation with anyone interested in learning more about some metrics about which they are unfamiliar. Learning is one of my favorite things to do in the world and it excites me to help someone accomplish that for themselves.

 

JH,

 

I tried to send you something pertaining to that Sabermetrics course, but couldn't because you have me blocked.

       
       


Oops. Try it now...