Scoring question

[SoCal_Dad]

Since you have no idea what any teams look like if presented the same way, how can you categorize our team as having depth? For all you know, we stunk compared to many other teams.

Pardon me for portraying your team in a positive manner. Regardless, the topic was determining whether the stats by position showed they suckered teams by manipulating the batting order.

Since you’ve never seen it, you have no idea as to its usefulness, do you. I can produce several way to look at batting positions. No one looks at them every single day, but periodically the coaches get together and analyze them all to see if they can tweak the lineup a bit to maximize its potential. I suppose if you’ve never seen anyone doing that, you could take the position that it isn’t useful beyond “satisfying a curiosity”, but others would categorize it more as “ignorance is bliss”.

I don't have to see other people do something to make a judgement about whether it is a good idea or not. I can evaluate the usefulness of stats based on a strong math background in probability, statistics and sample surveys. I seriously question the validity of using the reports shown to "tweak the lineup" for several reasons.

1. The stats for each position are a blend of multiple players and less useful the more they tweak the lineup.

2. Using stats based on a small sample size to make decisions during the season is risky because the distribution of outcomes is not uniform (e.g. streaks).

3. Good coaches know their players well enough to recognize when they should make adjustments because a player is struggling or on a tear.

4. "Past Results Do Not Guarantee Future Performance." This goes both ways - especially with small sample sizes. Poor stats for a good hitter that's hitting the ball well, but not getting their usual results, don't indicate what they're likely to accomplish going forward. Same with good stats for a mediocre hitter that got better than normal results for some weak balls in play.

That’s why they’re the largest purchasers and users of scoring software per capita in the world.

Per capita? Please provide some data to back up that claim.

 

[Greenmonsters]

Nobody knew who Bill James was when he worked as a night watchman.

 

[Scorekeeper]

Pardon me for portraying your team in a positive manner. Regardless, the topic was determining whether the stats by position showed they suckered teams by manipulating the batting order.

The reason I commented was you made a huge generalization without understanding what it was you were seeing. Sometimes people don’t realize just how little difference there is between players and teams, especially when you’re measurements are in hundredths or thousandths.

When did the TOPIC change to determining whether the stats by position showed they suckered teams by manipulating the batting order?

I don't have to see other people do something to make a judgement about whether it is a good idea or not. I can evaluate the usefulness of stats based on a strong math background in probability, statistics and sample surveys. I seriously question the validity of using the reports shown to "tweak the lineup" for several reasons.

1. The stats for each position are a blend of multiple players and less useful the more they tweak the lineup.

2. Using stats based on a small sample size to make decisions during the season is risky because the distribution of outcomes is not uniform (e.g. streaks).

3. Good coaches know their players well enough to recognize when they should make adjustments because a player is struggling or on a tear.

4. "Past Results Do Not Guarantee Future Performance." This goes both ways - especially with small sample sizes. Poor stats for a good hitter that's hitting the ball well, but not getting their usual results, don't indicate what they're likely to accomplish going forward. Same with good stats for a mediocre hitter that got better than normal results for some weak balls in play.

Well, how many times do you have detailed information about specific opponent players and make decisions accordingly, as opposed to just assuming the lower the BPOS the worse hitters are?

Yes, making decisions on small samples is riskier than making them based on large ones. However, making decisions on small samples is far better than guessing or flipping a coin.

Yes, good coaches do know their players, but it’s been proven many times how fallible human perception is compared to what’s actually happing. That’s why using real numbers is always at least a benefit if the coach is smart enough to know how to use them.

I agree, “Past Results Do Not Guarantee Future Performance”, but who said anything about predicting anything? I you assume the worst hitter in my lineup is the #9 hitter but I show you real numbers that prove that to be false and you don’t use them, does that make you the smartest coach in the world?

Per capita? Please provide some data to back up that claim.

I’m going by what the VP of marketing for GC said during a seminar. If you don’t like what he said, take it up with him.

 

[SoCal_Dad]

When did the TOPIC change to determining whether the stats by position showed they suckered teams by manipulating the batting order?

It didn't change - see below.

I saw/see relatively consistent numbers 1 thru 9 without any evidence of suckering teams. I can't tell whether there is a high correlation between those numbers and the hitters (i.e. depth) or there was more variation among the players and the 1-9 consistency was the result of moving them around in the order (i.e. suckering).

If you really want, I can print off all 16 pages of the individual breakdowns for each BPOS, and you can find your own correlations, if any.

That would be a bad approach. If the player stats for the regulars were in similar ranges like the stats by position (e.g. OBP .433-.495), you had depth from top to bottom.

Since you have no idea what any teams look like if presented the same way, how can you categorize our team as having depth? For all you know, we stunk compared to many other teams.

Pardon me for portraying your team in a positive manner. Regardless, the topic was determining whether the stats by position showed they suckered teams by manipulating the batting order.

The reason I commented was you made a huge generalization without understanding what it was you were seeing.

That "generalization" was IF your batters' numbers were in a similar range (see bold above). I based it on your past posts about the team and the ranges I've seen in the past for numerous teams. You haven't shown it's not accurate.

 

[FP26]

I take that to mean you believe QABs provide a better representation of a hitter than any other metric for small samples, but not for large samples. Then when you say: As the sample size increase, most if not all metrics improve in accuracy and validity That’s what confuses me. Why aren’t QABs equally representative later on?

Wow! That's not even close to what I was thinking... To keep it simple, let's just say that I think any stat utilizing a small sample size is unreliable (again, that's IMO). If all you have is a small sample, you need to utilize other indicators in order to help standardize the statistics. I see QABs as one of those supporting statistics for BA. I prefer larger sample sizes for all statistics. It simply isn't an option all the time.

NOTE: I never stated that QABs lose validity with a larger sample size.

 

[SoCal_Dad]

The topic of this portion of our discussion was the usefulness of your BPos reports.

Since you SoCal folks do everything so much better than anyone else, what does your team’s breakdown by BPos look like? I’ve never seen anyone else do that in amateur ball of any kind, but I’d be interested in seeing if we’re an aberration.

I haven't seen it and am not aware of anyone using it (e.g if iScore and/or GameChanger produces it). Data can be crunched and presented in an endless array of reports, but it's only information if it is useful beyond satisfying a curiosity.

Since you’ve never seen it, you have no idea as to its usefulness, do you. I can produce several way to look at batting positions. No one looks at them every single day, but periodically the coaches get together and analyze them all to see if they can tweak the lineup a bit to maximize its potential. I suppose if you’ve never seen anyone doing that, you could take the position that it isn’t useful beyond “satisfying a curiosity”, but others would categorize it more as “ignorance is bliss”.

I don't have to see other people do something to make a judgement about whether it is a good idea or not. I can evaluate the usefulness of stats based on a strong math background in probability, statistics and sample surveys. I seriously question the validity of using the reports shown to "tweak the lineup" for several reasons.

1. The stats for each position are a blend of multiple players and less useful the more they tweak the lineup.

2. Using stats based on a small sample size to make decisions during the season is risky because the distribution of outcomes is not uniform (e.g. streaks).

3. Good coaches know their players well enough to recognize when they should make adjustments because a player is struggling or on a tear.

4. "Past Results Do Not Guarantee Future Performance." This goes both ways - especially with small sample sizes. Poor stats for a good hitter that's hitting the ball well, but not getting their usual results, don't indicate what they're likely to accomplish going forward. Same with good stats for a mediocre hitter that got better than normal results for some weak balls in play.

Well, how many times do you have detailed information about specific opponent players and make decisions accordingly, as opposed to just assuming the lower the BPOS the worse hitters are?
How does this relate to your BPos reports?

Yes, making decisions on small samples is riskier than making them based on large ones. However, making decisions on small samples is far better than guessing or flipping a coin.
Here you advocate making decisions based on past results and 2 paragraphs later you say they're not being used to predict anything. Deciding on a line-up is making a prediction on what will work best for that game. Smart coaches know the limitations of the stats and make their decisions based on their knowledge of their players formed by their observations and reconciled with player-oriented stats.

Yes, good coaches do know their players, but it’s been proven many times how fallible human perception is compared to what’s actually happing. That’s why using real numbers is always at least a benefit if the coach is smart enough to know how to use them.
Key is knowing how stats should be used. As I described previously (see #'s 2-4 above), stats have limitations. Smart coaches use player-oriented stats as a feedback loop and reconcile any discrepancies with their evaluation of how the player is doing. Sometimes they adjust their evaluation and sometimes they identify where the stats are misleading.

I agree, “Past Results Do Not Guarantee Future Performance”, but who said anything about predicting anything?
As I stated earlier, a line-up is a prediction of what will work best for that game.

I you assume the worst hitter in my lineup is the #9 hitter but I show you real numbers that prove that to be false and you don’t use them, does that make you the smartest coach in the world?
I didn't assume anything about the hitter(s) in the #9 spot or the coach's strategy for putting them there. Some accomplished MLB managers, NL included, don't put their weakest hitter in the #9 hole.

You didn't address my concern about the BPos reports reflecting a mixture of players in positions and you haven't shown any benefit of using them over player-oriented ones.

 

[CoogansBluff]

With the advent of Scorekeeper, I have the feeling that these formerly innocuous ''How should I score this?" threads are going to be like some of those page 20-page "Critique my DD's swing'' threads.

Not that there's anything wrong with that.

 

[Scorekeeper]

Wow! That's not even close to what I was thinking... To keep it simple, let's just say that I think any stat utilizing a small sample size is unreliable (again, that's IMO). If all you have is a small sample, you need to utilize other indicators in order to help standardize the statistics. I see QABs as one of those supporting statistics for BA. I prefer larger sample sizes for all statistics. It simply isn't an option all the time.

NOTE: I never stated that QABs lose validity with a larger sample size.

If you don’t think QABs lose va\slidity as the sample size increases, why only then would you use other metrics? That was what stumped me.

BTW, I too PREFER larger to smaller sample sizes, but unfortunately that isn’t very possible in amateur BB or SB. Here’s the number of minimum sample sizes for some different offensive metrics. I don’t know about SB, but I’ve never seen or heard of any HS BB player getting 910 ABs to make for a stable BA. Heck, if every AB for an entire year were counted, including all HS, tournament, showcase, and all other venues were counted, I doubt any player would get that many. And no teams I know of show “lifetime” numbers for players because they don’t mix seasons or venues.

http://www.infosports.com/scorekeeper/images/sample.pdf

The result is, you go with what you’ve got because it’s better than guessing or depending on someone’s “gut”, or perception. As for whether a QAB is a good way to support other statistics, personally I don’t, but then again I don’t care what process anyone else uses to manage their team. If you want to use them, that’s fine by me.

 

[Scorekeeper]

With the advent of Scorekeeper, I have the feeling that these formerly innocuous ''How should I score this?" threads are going to be like some of those page 20-page "Critique my DD's swing'' threads.

Not that there's anything wrong with that.

It’s only because I’ve been doing this stuff for so long and know how difficult and involved they can get that I take issue with people trying to make things simple. At least with those other kinds of threads there’s usually a vid or pictures to look at.

When someone asks a question like the OP, I think it’s incumbent on those with some knowledge about the subject, to help the person asking the question frame it in a way to get the best possible answer without everyone trying to guess at things like what “hard grounder to the 3B” means. It’s only a matter if trying to help others get the “best” answers rather than a bunch of guesses.

 

[FP26]

My responses in bold....

If you don’t think QABs lose va\slidity as the sample size increases, why only then would you use other metrics? That was what stumped me. I use other metrics all the time, but with smaller sample sizes comes a larger margin of error, so other metrics become even more important. To hopefully prevent further stumpification, stating "more important" does not imply "no importance" at other times.

BTW, I too PREFER larger to smaller sample sizes, but unfortunately that isn’t very possible in amateur BB or SB. Here’s the number of minimum sample sizes for some different offensive metrics. I don’t know about SB, but I’ve never seen or heard of any HS BB player getting 910 ABs to make for a stable BA. Heck, if every AB for an entire year were counted, including all HS, tournament, showcase, and all other venues were counted, I doubt any player would get that many. And no teams I know of show “lifetime” numbers for players because they don’t mix seasons or venues. I agree

http://www.infosports.com/scorekeeper/images/sample.pdf

The result is, you go with what you’ve got because it’s better than guessing or depending on someone’s “gut”, or perception. As for whether a QAB is a good way to support other statistics, personally I don’t, but then again I don’t care what process anyone else uses to manage their team. If you want to use them, that’s fine by me.

I agree, mostly. QAB is a metric I like to keep, but it is not among the primary metrics I use. But if I am reviewing one weekend (15-20 at bats) and two players are both 5 for 16, I like to look at other things to differentiate. One weekend isn't going to change any long term plans, but it is part of the review process.

What stumps me is that you seemed to be stumped by my posts, but I am basically agreeing with 95% of what you are saying. I guess I don't explain myself too well. oh well...