The Big 'Do Averages make a difference to youth pull' Survey V2

[James018]

Batting average 55.16, current 12, potential 37.
Bowling average 28.07, current 17, potential 34.
Fielding average 0.28, current 3, potential 6.
Characteristics: Precocious, Opener

[James018]

Batting average 28.01, current 4, potential 28.
Bowling average 17.97, current 12, potential 36.
Fielding average 3.87 (keeper), current 5, potential 16.
Characteristics: Death bowler

[James018]

Batting average 33.08, current 7, potential 30.
Bowling average 26.69, current 12, potential 39.
Fielding average 1.48 (keeper), current 4, potential 13.
Characteristics: Late bloomer

[everett]

Thanks very much! The purpose of listing a large number of youth pulls is that, eventually, I will be able to do an unbiased statistical analysis on them to determine whether and (perhaps more importantly) how much difference the averages make. That will be more reliable than simply making guesses based on what we have observed, which are sometimes inaccurate.

Having said that, in my time tracking my team's youth pulls, I've noticed what I think is a very major correlation between players' averages and their current skills, but not their potentials. If I had to guess, I would say that averages only indirectly affect potential, in that a player's potential is always more than their current skill, so higher current skill means a higher minimum potential.

But we'll see what the numbers tell when I have time to analyse them properly.

Nicely done collating this data. I've had a half hour look at the data in this post, I make snap business decisions from data as a day job so thought this would be a fun challenge

Taking square bears data. I can see the following:

The average potential of the dataset is 30.10
The average currskill is 9.67
The max currskill seems to be 25, and min 0
Taking a histogram against potential shows the data approximates the normal distribution with a slight right skew. Thus the chance of having an aweful result is double that of having a brilliant result (potential 14-19 - 18 instances; potential 43-48 - 9 instances);

Sorting the dataset by potential, and taking quartiles (where progress = curr skill/potential):

Quartile 1: Bat Ave = 22.6; curr skill 5.81; potential 20.6; progress 27.8%; stdev progress 18.4%
Quartile 2: Bat Ave = 33.3; curr skill 9.64; potential 27.8; progress 34.7%; stdev progress 17.8%
Quartile 3: Bat Ave = 34.4; curr skill 10.4; potential 32.7; progress 31.8%; stdev progress 17.9%
Quartile 4: Bat Ave = 41.0; curr skill 12.8; potential 39.4; progress 32.9%; stdev progress 14.5%

The give-away was that the "progress" statistic in the 4 different quartiles approximated the mean with no clear correlation. A larger data set should show a reversion to the mean and my hunch is that intrinsic progress lies within the range of 30-33.33%

The deviation on the quartiles was also very similar on 3 of the 4 datasets (approx 18%), while the 4th quartile reflects a likely ceiling of 25 current skill , while no such bound applies on the lower 3 quartiles

Sorting the dataset by batting average, and dividing into quartiles shows:
Quartile 1: Bat Ave 14.9; Mean potential 25.6; St Dev Potential 7.8; Mean current 5.2; St Dev Current 3.5; Number observances >40 - 3
Quartile 2: Bat Ave 22.8; Mean potential 29.3; St Dev Pot 6.9; Mean current 6.2; St Dev Current 3.2; Number observances >40 - 4
Quartile 3: Bat Ave 39.3; Mean potential 31.5; St Dev Pot 6.1; Mean current 11.0; St Dev Current 4.1; Number observances >40 - 6
Quartile 4: Bat Ave 54.4; Mean potential 34; St Dev Pot 5.7; Mean current 16.0; St Dev Current 3.6; Number observances >40 - 6

This reflects James conclusion between a high correlation between average and current skill; On the other hand potential has an inverse relationship between batting average and standard deviation of potential.

In conclusion. Potential roughly approximates the normal distribution positively skewed for poorer outcomes. Current skills is roughly a third of potential with high variance. Max current skill is likely 25. Quartiles show aggregate trends between potential / bat average / current skill. Pulling a batter with a higher batting average produces a better batter but a more reliable result (ie highly likely to be range of 27-39), pulling a lower batting average produces a more variable result. As a general rule, I will be pulling batters with very high batting averages, or very low bowling averages (the top 10%) otherwise I'll pull based on overall scouting information (more impressive balanced scouting information; or bowling type I'm favouring)

[James018]

Thanks Everett! That's exactly the kind of analysis I would have liked to do but never found the time. You obviously have a great deal of knowledge and experience that I don't: what took you half an hour likely would have taken me an entire day

The most interesting thing for mine is that there seems to be a fair degree of correlation between batting average and potential, which I didn't expect: I thought there'd be minimal correlation if any at all, based on my anecdotal experience. Probably what fooled me is the smaller standard deviation which meant that super-high potentials are probably only around equally likely for all averages - but lower potentials are much more likely for worse averages than better ones. To use your quartiles for batting averages and the (very rough) assumption of normal distribution for each quartile, we'd expect that:

Quartile 1: Bat Ave 14.9; 95% of players have potential between 10 and 41.2
Quartile 2: Bat Ave 22.8; 95% of players have potential between 15.5 and 43.1
Quartile 3: Bat Ave 39.3; 95% of players have potential between 19.3 and 43.7
Quartile 4: Bat Ave 54.4; 95% of players have potential between 22.6 and 45.4

Considering it's only the best few players we're really interested in (those are the players we'd keep or make some decent amount of money from selling), it's probably that upper number we're more interested in, in each case. There's still a correlation between that and batting average but very minor. The higher initial skill, meaning a player is more likely to get closer to potential, probably makes more of a difference. Personally, I'll still be picking players with the highest average, but I can definitely see how someone might decide that getting the right style of player might be more important.

I would be interested in the same kind of analysis being done on bowling and fielding numbers, mind you. Bowling primarily to compare them against the batting numbers (e.g. is an average of 50 with the bat equivalent to 10 with the ball? Better? Worse?) and fielding because I've anecdotally not noticed any correlation between fielding averages and skill/potential and would like to see if it does exist.

[James018]

Batting average 23.99, current 7, potential 35.
Bowling average 23.12, current 21, potential 38.
Fielding average 0.13, current 4, potential 6.
Characteristics: Quick runner, Iceman

[James018]

Batting average 18.53, current 5, potential 32.
Bowling average 22.91, current 24, potential 48.
Fielding average 0.56, current 3, potential 14.
Characteristics: Precocious

[James018]

(Missed my youth last week.)

Batting average 7.74, current 6, potential 20.
Bowling average 19.15, current 10, potential 37.
Fielding average 0.28, current 5, potential 11.
Characteristics: New ball spinner

[MH713]

I just got a guy with 58.something batting average he has 14 current 46 potential, he had about 24ish bowling average 18 current, 36 potential,
occasional keeper with 0.96 catches per, 3/7 fielding, 2/16 keeping

[James018]

I just got a guy with 58.something batting average he has 14 current 46 potential, he had about 24ish bowling average 18 current, 36 potential,
occasional keeper with 0.96 catches per, 3/7 fielding, 2/16 keeping

While others' data is appreciated, it only really works if people do it every week. Otherwise if people only do it sometimes, it's more likely they'll remember to do it when they get a particularly good/bad/unusual player, which makes it look like those players happen more often than they really do. Not great for statistical analysis.

Anyway -

Batting average 17.54, current 5, potential 25.
Bowling average 40.90, current 10, potential 38.
Fielding average 0.31, current 3, potential 7.
Characteristics: Slow starter