125-year test sim result vs. default league total

[Jayzone]

All right, I have been setting up and few leagues (with few modified settings but default league total) and thinking about using one of them as my solo league. So, I ran the 20-year sim for them just to build up history before I dive in. Curiously, I found that the year-end league stats total seems to be... slightly offense inflated - particularly in the HR department. And this happens to all the leagues that I ran. Finding this strange, I decided to run a test league with mostly deafult settings and see what happens after running 100+ years of sims.

Here are the League settings:

2 sub leagues - one DH, one without DH
20 teams in total, 10 teams each sub league
Five levels of minor leagues (R to AAA)
Most of settings are left at default (including injury at normal)

Changed setting:
Amateur draft - 20 rounds with 21 rounds created

Total sim duration: 125 years

Here is the default league total setting that I used (and you see it in OOTP):

League total setting (default)
AB 166714
H   43972
K   32884
2B   9014
3B    886
HR   4878
BB  16337
HBP  1672

The league total was basically from the real 2008 MLB league total stats. We can put them into the context of per team per game (divide by 30 then divide by 162).

Tranlation into per team per game
AB    34.30329
H     9.047737
K     6.766255
1B    6.006996
2B    1.854733
3B    0.182305
2B+3B 2.037037
HR    1.003704
BB    3.361523
HBP   0.344033

Here is the result from the 125-year sim using the above mentioned test league.

125-year sim result (per team per game)
AB H K 1B 2B 3B (2B+3B) HR BB HBP
AB    34.80601
H     9.24356
K     6.874153
1B    6.06883
2B    1.80356
3B    0.234207
2B+3B 2.037768
HR    1.136963
BB    3.356306
HBP   0.45797

It would probably be easier just to put them into percentages compared to the league total setting (what we are "supposed" to get if left on default; according to the manual). So, here:

Put this into the context of percentages (Sim result/league total %)
AB H K 1B 2B 3B (2B+3B) HR BB HBP
AB    101.4655
H     102.1643
K     101.5946
1B    101.0294
2B     97.241
3B    128.4704
2B+3B 100.0359
HR    113.2768
BB     99.84482
HBP   133.1182

From the look of it, it seems that most of the sim results are very comparable with the league total settings - except 2B, 3B, HR, and HBP.

Doubles and triples are actually in line with league total setting if you compare the total of the two. This is because the difference between 2B & 3B is subject to runner speed, defense, park factors, and of course possibly relatively small sample size. Besides that, there is little distinction for where/how the balls was hit between the two. Since the total of 2B+3B is comparable between sim results and league total setting, I'd consider the result as accepatable but maybe improved with further tweaks.

HR and HBP are troublesome, being quite a bit off the mark. The HR total is off by an average of 21.588 HRs per team per season. And HBP is off by 18.4579. And this is pretty much in line with what I have seen with my other leagues. I do wonder why is this the case as with default settings, the league resulting stats should match up with league total according toe the manual. Even though park factors are random as per default setting, it is simple not enough to (and shouldn't) account for the difference. I will check on the park factors once I am home today but I don't believe the park factor will induce an overall HR inflation of 13.3%. (Edit: I have already checked. The league average park factor is about 1. So, no, park factor is not an influence here.) Plus, I have also seen this HR inflation in my other leagues. Of course, the reason may also be how AI is assemble the team (putting HR hitters in more?), possible league wide movement/power mismatch between pitcher and batters (developlment?), or the league may just not taliored to the league total in HR department - unlike the other stats. Or maybe the power ability rating spread is different than other ratings (league average PWR rating is higher than say, CON, EYE, etc. ?).

Of course, the solution maybe just to put in league total modifier of say 0.87 or so. Sure, this would bring the resultant sim result in line and the overall pitching performance would improve, but such solution will severely compress individual player HR output - as a previously 50 HR player would now become 43.5 HR player. And no, I don't want to see that happen either. Now, I am wondering what can be done to tweak the total in order to bring the ficitonal league output in line with what it is supposed to produce. Both on the league level and also individual level. Or maybe this is only something can be done by Markus.

[Jayzone]

Oh, and also, here is a graph of league HR sim results from year 1 to year 125.

The Blue is from the entire league. The Green is from the DH sub-league. The Red is from the no-DH sub-league.

The league total setting is rated at 162.6 HRs per team per 162-game season. As you can see, the majority of the seasons has a total much higher than that - including the supposely weaker no-DH sub-league.

[atlbrave1]

I have found similar issues w/ smaller samples (typically 20 - 25 years). I beleive the issue has to be w/ the AI or the player creation/development, since it occurs in both the default league and also in a league whose totals I modified to recreate 80's stat output.

Unfortunately, I don't know of any solution from the user end to fix this problem -- as mentioned, moving the PCM's around messes w/ the individual outputs, which for the most part I have found to be better than in prior year's. If it is the AI, there is nothing the user can do to stop it from doing its thing. It may be correctly evaluating players (i.e. selecting players who produce the most runs), but this process does seem to skew the league totals as time progresses.

I am interested if there is any solution to this problem, but have moved forward under the assumption that such solution will have to come from Markus and not the user.

[PhillieFever]

Unfortunately,park factors can absolutely cause a 13% increase in the number of homers.In my tests during beta,I found the ballpark factors to be quite potent in increasing offense,but not potent enough in decreasing it which unfortunately causes what you are seeing.

[atlbrave1]

Perhaps I was not concise enough in my post -- What I see occurring is that as time goes on, the runs scored per game increase. Here is an examples from a 30+ test league I have run:

Settings: MAL, Standard - No modifications
League Totals (Standard)
AB: 166714
H: 43972
2B: 9014
3B: 886
HR: 4878
BB: 16337
HB: 1672
SO: 32884

Total League HR's:
Year 1: 4466
Year 5: 5403
Year 15: 5999
Year 25: 6081

Runs Scored Per Game:
Year 1: 5.08
Year 5: 5.18
Year 15: 5.32
Year 24: 5.43

All other stats stay closer to the league totals (here is quick poorly formatted summary (Yr1, Yr5, Yr15, Yr25)
2B: 9828, 9452, 9307, 9196 (Higher, but declining through time)
3B: 1225, 1143, 1201, 1133 (These are obviously high (but somewhat consistent)
BB: 16414, 16518, 16826, 17097
SO: 30868, 31472, 33967, 32935

I have run other tests with similar results --- Over time, More HR's are hit -- My theory is that the AI values walks and HR's over other skills (I'm not necessarily disagreeing w/ it, just stating my thought of what the AI is doing) and therefore the player mix gets skewed as time moves on.

I'm all for ebbs and flows, but it seems from the few 25 year samples I have run, the same escalation in HR's and runs occurs. I have yet to see a 25 year run where runs decreased in the same manner, which is somewhat bothersome.

Just wondering if anyone else is seeing the same time of escalation, or if I need to run 100 25 year tests to confirm (assuming I could find a babysitter for the remainder of the week)

[Jayzone]

Originally Posted by PhillieFever

Unfortunately,park factors can absolutely cause a 13% increase in the number of homers.In my tests during beta,I found the ballpark factors to be quite potent in increasing offense,but not potent enough in decreasing it which unfortunately causes what you are seeing.

I just checked my league files and I must say otherwise. No, the park factor is NOT a reason in the HR inflation here. All the park factors (for the ML team) have a (almost) linear spread ranging from 0.83/0.924~1.1 with 0.83 as the outliner. In any case, the average value of the whole league ends up being 1.00531 - which is consistent with the value that it should be. This means the league in its entirety is neutral park environment. So, the HR inflation has to come from somewhere else other than park factors.

[Jayzone]

As for the R/G, I can't say that I am seeing that trend either in this test sim or other leagues that I have. A short period of increase or decrease, yes, but never a continous increasing/decreasing trend over a very long period of time.

The attached graph is what I have for this test sim. The closest thing that I am seeing is the run trend between year 70 to year 91 (a general trend of increasing upward). However, after that, the league stabilize backtoward its average R/G value again.

Note: the average value of R/G for this 125-year test sim is 5.0344 with a max of 5.323148 and a min of 4.700349. The first year value is 5.188889.

[atlbrave1]

Jayzone-

Thanks for the info. I guess I'll run a few more tests --- maybe I was just running into anomalies w/ respect to player creation. Or maybe Bud Selig hadn't cracked down on steroids yet.