Note: more detailed explainations about this study are in my previous posts
I have an update for my little 'player ability progression study'. This update is about the break-out age that I talked about last time and this is for the 1982 draft class (I am still working on second draft class (1983) in analysis.). First,
disclaimer: small sample and intermediate study result. Since it's a small sample, I won't write much in terms of analysis.
I thought about how to define "break-out age" or more precisely "potential fullfillment age". I set a few criterion for both % A/P and % A/Pi ratio. I try to map the distribution of break-out age for the specific criterion. (i.e. Player X achieve 90% A/P ratio at the age of Y, etc.) So, in the end, there will be an age ditribution table of players for specific critierion. I did this is both number of players and percentile. I also calculate the max/min/average/standard deviation for all the distribution curves. So, we will have players avhieve %A/P=80% at the average age of Y. One thing to note though,
NOT every player will statisfy the criteria. So, all this max/min/avg/stdev numbers are for
players who made the criterion. This is actually quite important, as this not-covering-every-sample-avilable tend to screw the result on average. This means that you can not just assume a player will avhieve certain A/P ratio criterion in a average of age X. This X is only meaningful if the player can actually make the criterion. So, for this reason, it's important to pay attentaion to "Never" column in the table. The "Never" column is for players that have never reach the specific critreion.
Another thing to note is that lower-roud/pick players tend to have higher intital %A/P ratio since their intial potential is low to begin with. In this draft class, for example, there are quite a few low-pick players who has high A/P ratio (oever 80%) but they are just projected to career minor leaguers (due to low potential). This intial higher %A/P also screw the distribution a little bit. Because of this and I wasn't too sure where to set the separation-line, I set the potential fullfillment criteria not just with one number but with a few (80%, 85%, 90%). So, I hope this broadening on the selection range will give you a better overall picture.
Besides %A/P ratio, I also did the %A/Pi ratio. Why? It's because that %A/P ratio does not tell the whole story. A player can have high % A/P but low % A/Pi due to potential drop. So, even if the player statisfy the specific %A/P criterion, I still don't think I should call this as a "break-out" year. It's just the player has reached the growth ceiling at that year. On the other hand, % A/Pi ratio gives a better picture on if the player truly has a break-out year or if the player has really fullfill the initial project at draft. The criterion range is even larger than %A/P (80%,85%,90%,110%,125%). The reason for this is that I want to see what age does the player exceed initial draft projection if at all. So, these are for "surprise players". The age for these over 100% criterion will be older because players need time for both potential growth and actual ability growth.
So, without further confusing the readers with my explainations. The analysis on potential fullfillment age for 1982 draft class is shown below.
Code:
"Break-out Age" or "Potential fullfillment Age"
% A/P ratio:
criterion vs age distribution (number of players; out of total of 84)
18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Never
80% 1 4 8 14 28 12 7 9 2 1 8
85% 1 2 5 11 13 9 14 10 5 2 1 11
90% 1 2 4 2 9 8 15 7 10 7 5 14
criterion vs age distribution (percentile)
18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Never
80% 1.19 4.76 9.52 16.67 33.33 14.29 8.33 10.71 2.38 1.19 9.52
85% 1.19 2.38 5.95 13.10 15.48 10.71 16.67 11.90 5.95 2.38 1.19 13.10
90% 1.19 2.38 4.76 2.38 10.71 9.52 17.86 8.33 11.90 8.33 5.95 16.67
Group potential fullfillment age stat for % A/P
80% 85% 90%
Max 31 31 31
Avg 25.3 26.0 27.1
Min 21 21 21
Std 1.93 2.09 2.41
% A/Pi ratio:
criterion vs age distribution (number of players; out of total of 84)
18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Never
80% 2 3 4 6 15 8 6 8 7 2 1 1 21
85% 1 4 1 2 15 7 4 6 6 4 2 1 31
90% 1 1 1 4 5 10 4 6 4 1 2 2 2 1 40
110% 1 2 1 3 3 2 1 1 70
125% 1 1 3 1 1 77
criterion vs age distribution (percentile)
18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Never
80% 2.38 3.57 4.76 7.14 17.86 9.52 7.14 9.52 8.33 2.38 1.19 1.19 25.00
85% 1.19 4.76 1.19 2.38 17.86 8.33 4.76 7.14 7.14 4.76 2.38 1.19 36.90
90% 1.19 1.19 1.19 4.76 5.95 11.90 4.76 7.14 4.76 1.19 2.38 2.38 2.38 1.19 47.62
110% 1.19 2.38 1.19 3.57 3.57 2.38 1.19 1.19 83.33
125% 1.19 1.19 3.57 1.19 1.19 91.67
Group potential fullfillment age stat for % A/Pi
80% 85% 90% 110% 125%
Max 33 34 34 32 34
Avg 26.0 26.5 27.2 28.4 30.1
Min 21 21 21 24 26
Std 2.48 2.66 2.96 2.13 2.48
research status: The second draft class analysis is slowly coming along. I am hoping that I can post the result/analysis after looking at five draft-classes. And perhaps 10,15 draft class analysis someday.
Thanks for reading. Observations/Suggestions are welcome.