Pitchers only throwing fastballs

[Bristolduke]

I'll toss this out there although it's a different stat. I created a bug report #UYU-445281 where fielding ratings were not being created with using Neutralized stats AND not having the .odb file. I saw this on my randomization files, where I was changing year but was leaving all the other .csv data as is.

I had assumed the .odb was being used to allow import history not to create ratings.

The difference I see in Steve's test vs the reported problem, is that TwinsFan is always starting players at their debut year. On the 58 and 60 tests Wynn and Sphan are being imported in the middle of their careers.

Marcus has indicaed one error in this area and that may not be the only one based upon the different scenarios reported. Some testing in this area after the patch would be prudent. I'll be glad to help in that area if needed.

[TwinsFan]

Originally Posted by Bristolduke

The difference I see in Steve's test vs the reported problem, is that TwinsFan is always starting players at their debut year. On the 58 and 60 tests Wynn and Sphan are being imported in the middle of their careers.

Exactly. I've run some tests and Sphan imports correctly if you import in the middle of his career. This error is mostly happening when you draft the player. Look at any draft class and you will find 10-15 pitchers that only throw a fastball.

[SteveP]

Originally Posted by TwinsFan

Exactly. I've run some tests and Sphan imports correctly if you import in the middle of his career. This error is mostly happening when you draft the player. Look at any draft class and you will find 10-15 pitchers that only throw a fastball.

Well, keep in mind that in my 1958 and 1960 leagues, there are pitchers who are debuting in those years. But I don't want to belabor that point, because of the many variables involved.

I can't imagine what would cause a malfunction in downloading the repertoire data in the DB for pitchers, when it is their debut year. If anyone else can think of why that might happen, it would certainly help the debugging process. There might be enough info here for Markus to decide to try duplicating the bug. OTOH, I've seen problems like this where it only seems to affect a small number of OOTPers, so he may not be able to duplicate it without more to go on.

[Bristolduke]

A quick analysis shows Ryan, Wynn and Spahn have a common thread as it relates to computing ratings.

Each pitcher had only a few innings in their first year, Ryan-3, Wynn-20.1 and Spahn-15.2. All of these are under the default adjust of 40 innings with Spahn and Ryan being under weaken value of 20. (I am assuming these settings aren't being changed)

Second, all three pitchers have no real stats the following year (Spahn actually has a gap for 4 years). This probably explains why actionjackson got different ratings on Ryan when neutralized was set, because those ratings fill in the gap years that exist from the real stats.

So if we were looking for a pattern to begin to analyze and debug, those conditions would suggest a starting point.

I'll try to run some analysis, if that will help.

[Bristolduke]

I did a run of 1950, but since I'm not sure what relationships I'm looking for, running off a doing a lot of analysis may be a waste of time. My assumptions are as follows:

1) Since pitches are contained in the Master.csv file and thus are not season variavles, that each pitcher should have a rating for each pitch he is rated in the file at time of import. Then based upon the historical settings, the pitch ratings and pitcher ratings would vary based upon what was being caluclated. Thus a 1 year recalc might look different than a career rating and no recalc.

2) A fastball rating is almost a given for any pitcher. If players get default fastball ratings of 1, a pitcher has to be at least a little better, certainly not worse.

3) when dealing with reduced innings and thus adjustments, the pitch ratings would have to be the same (e.g. unadjusted), because there is only one set of them. If the pitches are reduced in the adjustment, there must be a mechanism to reset them when the pitcher gets over the adjusted value. My assuption is that pitch ratings are not adjusted.


Parameters used.

I started a historical 1950 season without importing history. I left the parameters on default, which means current ratings on real-stats, potential on entire career, 3-year for pitching and fielding. Re-calc is set at 3-years for when I go forward.


Of the 230 pitchers imported, 25 had only fastball ratings and 26 had no fastball ratings. (each position player had fastball ratings).

It appears that if a pitcher has a fastball rating of 0 that is worse than a player with a null/blank rating since all position players are rated better than these pitchers.

It appears that all 26 pitchers with no fastball rating had a rating of 0 in the Master.cvs (I only spot checked the starters Blackwell, Burdette and Lemon).

6 of the 25 pitchers with only a fastball rating also had a fastball rating of 0 in the master.csv.

All 25 pitchers with only a fastball rating were rated in other pitches as well in the master.csv.

15 of the 25 pitchers with fastball only ratings pitched less than 10 innings in 1950.

Only one pitched more than 40 (the adjustment default), Steve Nagy-53.

Pitchers with only Fastball ratings

Kent Peterson
Bill Pierro
Jim Romano
Chris Van Cuyk
Sid Schacht
Phil Marchildon
Blix Donnelly
Hal Gregg
Al Aber
Ribs Raney
Ernie Johnson
Dick Littlefield
Ed Albrecht
Dick Donovan
Steve Nagy
Carlos Pascual
Joe Landrum
Les McCrabb
Frank Barrett
Mal Mallette
Gus Keriazakos
Rogelio Martinez
Gordie Mueller
Cot Deal
Charlie Cuellar


Is this on the right track with respect to helping isolate code issues? Or should something else be looked at as well or instead?