Ground balls and percentages

[Greenmonsters]

Now that’s a unique way of presenting stats! It took me a while to figger out what was going on, but I got a handle on it. Although I see nothing at all wrong with doing it that way, I think it would be difficult for someone trying to compare players because of the sorts, but as far as I can see its fine. Having said that, I can say that that’s the only place I’ve ever seen stat’s presented that way.

I have to say though, that I don’t see anything in those “per game” numbers that are “reported as # of outs normalized per 100 PAs”. Can you help me find that? It looks to me like the per game numbers are just the number of something divided by the number of games that player played.

Thanx

SoCal's on target. Not sure why the stats are presented that way. But when looking at FB and GO, I'd rather have spray charts with outs/hits and GB/FB/LD toggles ala Game Changer.

 

[Scorekeeper]

See the note - * = average based on 100 plate appearances

I guess I’m missing something simple here. In the Per game stats, from I can determine, all of the different stats aren’t per game. All of the columns labeled “X”/g are. They take the gross numbers and divide them by the number of games. The columns labeled *”X” are something different. Those are the one’s claiming to be based on 100 PAs.

The problem I’m having is, I’m not used to seeing numbers that have come from some algorithm, without seeing either that algorithm or all of the data points used in it. FI, the note says based on 100 Plate Appearances, but there’s no place that list them. So what someone like me has to do is to go to the Overall Stats and compute PAs on my own by adding up ABs, BB, HBP, Sac, and sacf. Then I can compute the factor they’re using to “normalize” for the 100 PAs. I’m not saying its right or wrong or good or bad, I’m just not used to seeing things presented that way.

When I try “normalize” things, rather than try to make everyone’s PAs equal and then show what the numbers would be, I’ll more likely just show things “Per PA” or “PAs Per”. To me its easier to understand, but its rally all in what ya get used to.

 

[Scorekeeper]

SoCal's on target. Not sure why the stats are presented that way. But when looking at FB and GO, I'd rather have spray charts with outs/hits and GB/FB/LD toggles ala Game Changer.

When I’ve generated spray charts, I’ve spent the extra time to break down fly balls into pop ups and fly balls. Granted it subjective, but since I’m the only who keeps score at our school, people can be pretty sure that there will be consistency in how things are scored.

And while I love spray charts, to me there are other ways to present those things that are a bit easier to comprehend. Here’s an example.

http://www.infosports.com/scorekeeper/images/gbfb1b.pdf

If you look at the last page, it becomes pretty easy to pick out the players who are most likely to hit say a line drive when they hit the ball. Not a better way perhaps, but different.

 

[Greenmonsters]

I guess I’m missing something simple here. In the Per game stats, from I can determine, all of the different stats aren’t per game. All of the columns labeled “X”/g are. They take the gross numbers and divide them by the number of games. The columns labeled *”X” are something different. Those are the one’s claiming to be based on 100 PAs.

The problem I’m having is, I’m not used to seeing numbers that have come from some algorithm, without seeing either that algorithm or all of the data points used in it. FI, the note says based on 100 Plate Appearances, but there’s no place that list them. So what someone like me has to do is to go to the Overall Stats and compute PAs on my own by adding up ABs, BB, HBP, Sac, and sacf. Then I can compute the factor they’re using to “normalize” for the 100 PAs. I’m not saying its right or wrong or good or bad, I’m just not used to seeing things presented that way.

When I try “normalize” things, rather than try to make everyone’s PAs equal and then show what the numbers would be, I’ll more likely just show things “Per PA” or “PAs Per”. To me its easier to understand, but its rally all in what ya get used to.

I actually like it normalized to 100 PAs as that's relatively close to a full season - easier for me to put in perspective than on a #/game basis where small differences in the numbers after the decimal point have a significant impact. Looking at the normalized projections, AB does not equal H+K+GO+FB. I believe SH and SAC are included but not listed, but that still doesn't add up. I believe ROE also is not included and unlike SH and SAC, its a stat that's not reported in any the other stat tabs on the conference website. But like you've said before, unless you know what goes into a stat (or an algorithm), you never really know whether its meaningful (or accurate)!

 

[SoCal_Dad]

I played around with the numbers and looked at a few box scores.
- The raw FB and GB numbers are in the Analysis Stats and labeled as Fly Out and Gnd Out. They do include SH and SF (verified w/ box scores).
- The per-100-PA values are truncated, not rounded (e.g. 85.52 is reported as 85). Either way, you can't reconcile using them because you need at least another digit of precision.
- ROE is not included nor reported, so that mucks up reconciling to AB and/or PA.

I don't like the per-game stats because they inherently favor starters over subs because they usually get more PA per game. Using 100-PA instead of PA actually just scales the decimal values to percentages (e.g. .85 becomes 85) and makes showing a decimal point unnecessary.

 

[Scorekeeper]

I actually like it normalized to 100 PAs as that's relatively close to a full season - easier for me to put in perspective than on a #/game basis where small differences in the numbers after the decimal point have a significant impact.

The neat thing about numbers is, as long as the math is done correctly to get to the final result, the final result is still the same. You like to see things extrapolated as though everyone had 100 PAs, I like to see it presented as relative to the number of PAs, and there are lots of folks who like to see it lots of other ways.

Looking at the normalized projections, AB does not equal H+K+GO+FB. I believe SH and SAC are included but not listed, but that still doesn't add up. I believe ROE also is not included and unlike SH and SAC, its a stat that's not reported in any the other stat tabs on the conference website.

For many years I entered ABs into the database, then computed PAs on the fly by adding BB, HBP, Sacf, SacH, and CI. What I found was, it was more of a PITA to enter the data because I had to do double duty entering the data. So I changed my program to allow me to enter PAs, then let it compute the AB field, and stored both fields. Believe it or not, just that little change saved me a lot of data entry time. MaxPreps has users enter ABs, but it doesn’t use PAs in any calculations. My guess is, those Liberty stats are entered the same way. The difference is, when they decide to us PAs, they probably then have to calculate them on the fly.

But like you've said before, unless you know what goes into a stat (or an algorithm), you never really know whether its meaningful (or accurate)!

Its really hard to be transparent when presenting the numbers the way Liberty does, i.e. the normal way. I do hitting, pitching, and fielding stats too, but present them as many different reports in each that show many different things, rather than one report. As I’ve said before, its all in what you get used to.