Age Distribution

[jarmenia]

I know its a small sample size but I decided to run the sample dump given to us today through a little query to see how the age of the players is distributed through the league. Here's what I've found for the batters:

Age	CountOfplayer_id
21	6
22	8
23	13
24	35
25	14
26	14
27	11
28	14
29	12
30	12
31	17
32	14
33	13
34	13
35	8
36	4
37	7
38	5
39	11
40	1
41	4
42	1

Granted is a small sample size as its just for the one year that was dumped, and only includes players with a batting record, but does this look right to somone in the know. Seems a bit skewed to my untrained eye.

[DrArbiter]

Originally Posted by jarmenia

Granted is a small sample size as its just for the one year that was dumped, and only includes players with a batting record, but does this look right to somone in the know. Seems a bit skewed to my untrained eye.

Not really. The mode is approximately in the right place (around 24), though the bump there is a little overlarge. Usually you see a bit more steady drop after 24.

And before someone points this out, yes, most players do peak around 27-29 or so. But remember there's a selection process going on, too, driven by the player development and movement rules. If by the time a guy is 24 or 25 and makes it to AAA, and maybe a late-season callup to the bigs, and he isn't particularly impressing, he becomes progressively less attractive to teams. Often, those guys kick around for a few years in AAA as filler or emergency backups, but most clubs will prefer to bring up and develop a younger guy.

[disposableheros]

i wonder what it looks like combined with the pitchers. they tend to be a tad bit on the older side (come up a little later, peak a little later, stay around a little later).

[Malleus Dei]

I don't like that bump at 24 much. I'd like to see a breakdown on all the members of that class.

[andymac]

Originally Posted by Malleus Dei

I don't like that bump at 24 much. I'd like to see a breakdown on all the members of that class.

The sample size is much too small to draw any conclusions from it.

[nestajones]

Originally Posted by DrArbiter

Not really. The mode is approximately in the right place (around 24), though the bump there is a little overlarge. Usually you see a bit more steady drop after 24.

NO. The mode age in the major leagues is absolutely NOT 24. The average age has been around 28-29 in recent times (this has changed throughout history), and I assume the mode (for whatever its worth) is in there too.

You can check average ages by team on baseball-reference.com, though those are not exact because they are weighted by playing time. (This might be more informative, depending on what you want to know.) Still, look at the roster of any team...90% of the players in the majors are older than 24, and probably only about 50-60% debut by that age.

Also look here: http://www.baseballmusings.com/archives/008682.php

In my view, this is important to get right. People can change it with modifiers, but the default should really echo actual aging patterns.

Of course, it should be noted that average age in the majors is not only the result of aging patterns, but also draft rules, 40-man roster rules, options, free agency, finances, and team strategy. It's easy to set the game rules to mimic MLB roster rules, but obviously perfecting the AI decision-making is tougher.

[jarmenia]

Originally Posted by Malleus Dei

I don't like that bump at 24 much. I'd like to see a breakdown on all the members of that class.

We can do more looking once we get a full dump or actually get the game. It just looked strange to me.

[rudel.dietrich]

The lack of anyone below 21 is a bit of a concern

On the other hand the numbers of players 37 and over are comforting to see. It is about dame time older players started hanging around

[jbmagic]

I think Skydog did a test on age distrubution before when he did his modifiers setting for ootp 6. He might still have what the averages should be.

Hopefully he can look at age distrubution for batters and pitchers before the final release.

It will be nice to have the default setting for ootp 2006 have good age distrubutions already. This way we don't need to mess with the modifiers unless we want something extreme.

[DrArbiter]

Originally Posted by nestajones

NO. The mode age in the major leagues is absolutely NOT 24. The average age has been around 28-29 in recent times (this has changed throughout history), and I assume the mode (for whatever its worth) is in there too.

Your error is in your assumption that the mode is about the same as the average. There's no reason to assume a symmetric distribution of ages -- in fact, given the way roster rules and the like work, a priori one should assume the distribution isn't symmetric.

Thought experiment to see why: if you had a 23-year-old and a 29-year-old of equal current skill, which would you release, and which would you keep on the roster?

You are right that the average age is around 28-29. But the modal age is usually 24 or 25, with a gradual but steady dropoff thereafter.

Before making that post earlier, I checked several years in the early-to-mid 80s (which I just happened to have handy), and the mode was always 24 or 25.

What seems odd about the distribution originally posted is the huge *drop* after 24, which messes up the average.

[nestajones]

Originally Posted by DrArbiter

Your error is in your assumption that the mode is about the same as the average. There's no reason to assume a symmetric distribution of ages -- in fact, given the way roster rules and the like work, a priori one should assume the distribution isn't symmetric.

I don't assume it's necessarily symmetric, but with a (basic) range of ages from 20-40, the distribution would have to be very screwy for the mean to be 28-29 and the mode to be 24.

Thought experiment to see why: if you had a 23-year-old and a 29-year-old of equal current skill, which would you release, and which would you keep on the roster?

I see what you're driving at, but I don't know how to answer that question. How has "equal current skill" been determined? That can never been known for sure in real life. Some teams will take the older guy as more of a "sure thing," or to let the younger guy get more minors seasoning.

You are right that the average age is around 28-29. But the modal age is usually 24 or 25, with a gradual but steady dropoff thereafter.

Before making that post earlier, I checked several years in the early-to-mid 80s (which I just happened to have handy), and the mode was always 24 or 25.

Can you post some data on that, because I find it extremely hard to believe. I don't have the time to calculate or look up the numbers right now, but from eyeballing some 2005 teams, there is no way the modal age is 24-25. The avg age was about a year less in the early 80s, but I really don't think it would have changed the modal age that much. I am open to data, however.

What seems odd about the distribution originally posted is the huge *drop* after 24, which messes up the average.

Well, it is a small sample of players, so I don't know how much to read into it. What kind of league did the data come from? A newly formed league? One 50 years old? It was so skewed to 24 year olds that it does bother me, however.

[DrArbiter]

Originally Posted by nestajones

I don't assume it's necessarily symmetric, but with a (basic) range of ages from 20-40, the distribution would have to be very screwy for the mean to be 28-29 and the mode to be 24.

Not at all. (Are you maybe thinking of the median, not the mode?) The general shape of the distribution is a fast ramp-up from 21 (there are usually rather few 21-year-olds) to 24, when the distribution peaks; then, there's a slow and gradual decrease thereafter. That sort of distribution can easily give you an average age of 27-28 while keeping the mode at 24.

I see what you're driving at, but I don't know how to answer that question. How has "equal current skill" been determined? That can never been known for sure in real life. Some teams will take the older guy as more of a "sure thing," or to let the younger guy get more minors seasoning.

Agreed in principle; certainly different people and teams feel differently. But, in my scenario, the 29-year-old is basically as good as he's going to get (probably), while the 23-year-old is likely to get much better. So the 29-year-old might or might not get another chance, but you know the 23-year-old, barring injury, will get his chance.

Can you post some data on that, because I find it extremely hard to believe. I don't have the time to calculate or look up the numbers right now, but from eyeballing some 2005 teams, there is no way the modal age is 24-25. The avg age was about a year less in the early 80s, but I really don't think it would have changed the modal age that much. I am open to data, however.

I did it for a few leagues yesterday in an idle minute by copying the all-players page of baseball-reference into Excel and using text-to-columns. Unfortunately, it crashed before I transcribed it, and I'm not that interested in doing it again. But it is simple enough to do, if someone wants.

[tysok]

I pulled up and sorted through the age info on 2 years of national league baseball, 1996 (because I already had it sorted) and 2005.

1996
315 batters, 268 pitchers
Average age for batters: 28.04
Average age for pitchers: 27.77
Batters
19 years old: 1
20: 4
21: 6
22: 9
23: 15
24: 27
25: 32
26: 29
27: 28
28: 30
29: 25
30: 27
31: 19
32: 18
33: 13
34: 14
35: 7
36: 2
37: 5
38: 1
39: 1
40: 0
41: 2
Pitchers
21: 7
22: 7
23: 16
24: 28
25: 26
26: 30
27: 27
28: 28
29: 23
30: 20
31: 16
32: 7
33: 8
34: 7
35: 6
36: 3
37: 2
38: 3
39: 1
40: 1
41: 1
42: 1

2005
342 batters, 341 pitchers
Average age of batters: 28.81
Average age of pitchers: 28.61
Batters
20: 1
21: 8
22: 12
23: 15
24: 24
25: 34
26: 31
27: 29
28: 25
29: 32
30: 14
31: 21
32: 18
33: 19
34: 14
35: 13
36: 8
37: 9
38: 9
39: 2
40: 3
41: 0
42: 0
43: 0
44: 0
45: 0
46: 1
Pitchers
20: 3
21: 8
22: 11
23: 21
24: 28
25: 22
26: 29
27: 38
28: 32
29: 24
30: 16
31: 17
32: 22
33: 16
34: 10
35: 14
36: 7
37: 8
38: 4
39: 7
40: 1
41: 0
42: 2
43: 0
44: 1

Make of that what you will.

Everything fixed as well as it can be... unless something stupid (like having only some of the pitchers in the batting lists but not all) it's accurate as can be.

[andymac]

Originally Posted by tysok

Make of that what you will. Keep in mind the pitchers seem to be mixed in with the batters, even if they never had an at bat. It was the same in the list for the AL and I didn't bother to weed them out.

All you should need to do is subtract the pitcher numbers from the batter numbers then right?

[tysok]

I suppose that's right... I'm far too lazy to think right now. But I'm ahead of schedule still so I'll edit that in a minute.

[tysok]

For some reason I thought it was 2 hours later... early parent teacher conferences suck. Anyhow, some players were listed twice (had the breakdown by team). I'll get this edited in about an hour.

[tysok]

All fixed. And I'm still ahead of schedule so I can take a minute to laugh at my mistakes...

Hahaha...

Okay time to go...

[jbmagic]

Originally Posted by tysok

All fixed. And I'm still ahead of schedule so I can take a minute to laugh at my mistakes...

Hahaha...

Okay time to go...

You should pass that along to Markus.

[Marc Duffy]

Yes, please do

[tysok]

Not sure what you guys mean.... anyhow... I thought I'd do 1980 today just cuz someone mentioned it might have been different earlier in baseball history. It's kidna funny... baseball is getting older it would seem. You could see an average age of 50 by 2090 if the average age continues at these paces.

1980
248 batters, 183 pitchers
Average age of hitters: 27.75
Average age of pitchers: 27.56
This is hitters - pitchers
19: 0 - 2
20: 4 - 0
21: 3 - 5
22: 13 - 8
23: 17 - 17
24: 28 - 15
25: 21 - 22
26: 20 - 19
27: 23 - 17
28: 19 - 13
29: 21 - 9
30: 15 - 15
31: 16 - 9
32: 20 - 5
33: 6 - 7
34: 9 - 3
35: 3 - 9
36: 3 - 0
37: 1 - 4
38: 1 - 1
39: 1 - 0
40: 1 - 1
41: 0 - 2
42: 2 - 0
43: 1 - 0