Update: Rating Translator/Expected Performance Scout v0.93b is now available!

[deckholm]

The link on jazzrack's site doesn't work for me. Does anybody have another link where the rating translator is located? Thanks.

[Jayzone]

Sorry for the late reply, I have been busy with writing thesis and thus only check out the board occasionally.

Originally Posted by genghisdunn

I'm just getting into using this thing, and it all works great. My question is: is there a way to generate the projections for every player automatically, perhaps into some kind of spreadsheet? I track the entire league right now, and it's going to be awfully time-consuming to go through each player 1 by 1, scout their performance, and then enter that info into a spreadsheet so that I can manipulate it. Is there something I'm missing in the program itself that will generate this type of output file? Or has someone written a program that will do that?

genghisdunn,

The short answer is no. I was planning on expanding the functionality of rating translator, which would've included function of the mass translation onto a csv file. Unfortunately, I didn't get to finish that and I probably won't have time for anything other than my thesis for the forseeable future. So, v0.93b will have to make do. Really sorry about that.

p.s. Oh, I also don't recall anybody writing a utility utilizing the output information from the rating translator (except the included GUI interface by Eriq Jaffe, of course).

Originally Posted by deckholm

The link on jazzrack's site doesn't work for me. Does anybody have another link where the rating translator is located? Thanks.

deckholm,

Not sure what is happening to jazzrack's site. I just checked the download page and it seem the download function is down for all files. Well, I don't know if the newest rating translator is hosted somewhere else, but the best and fastest solution that you could probably do is to contact jazzrack via email or PM him about the download problem on his site. I will also contact him when I have the time, but probably not imdediately.

Contact info from jazzrack's download page: jazzrack@dawnstarv.us

All right, back to thesis.

[Rolen2k2]

okay...so im using this thing, and everything works perfectly until i hit scout...then after it attempts to load something, no output window loads and the program shuts down

whats going on?

[Jayzone]

Originally Posted by Rolen2k2

okay...so im using this thing, and everything works perfectly until i hit scout...then after it attempts to load something, no output window loads and the program shuts down

whats going on?

Sorry for the lateness of the reply, still working on my thesis.

Hmm.... without the detail of what is exactly happening, I can't really say for sure what exactly is the cause of the problem. However, I could offer my best guess on the source of the problem. It sounds like that the CSV data file format is not 'normal' so the read_line() function probably has some trouble interpretating the data in a line. I think there is probably some weird characters in some of the text columns in players.csv file and the trouble-causing column will probably be either LName, FName, NName, Hometown, or HomeState. Sometimes, when part of weird character byte data is being read without context, it will be interpreted as EOL character and this would likely to cause problems. One way to check this is to open the players.csv with the notepad application and check to see if any line is being cut in half prematurely. For example:

correct line format:

...,"OOTP2006 beta-tester",....

incorrect line format:

....,"OOTP2006 be
ta-tester",...

As it can be clearly seen, the data of this line in CSV file is being intepreted wrongly. Thus, one complete line is interpreted as two lines. If this is indeed the issue, then the solution will be to use the notepad getting rid of the weird characters or to use any spreadsheet program to delete the troublesome column once and for all. As long as the players.csv data format is normal, the program should work.

Hope this does help a bit. All right, time for me to go back to work.

p.s. By the way, the CLB download link is working again now. So there shouldn't be any problem with that.

[Jadetorc]

I've seen several threads on the creation, and while this question has been asked before, I haven't seen an answer: What are the formulas for translating PITCHING stats into ratings? Ignoring league totals, ignoring park factors, ignoring defense, L/R splits, endurance and injury. I'm sure they aren't linear, so how they are interrelated is the important issue. I've seen and use calculations for batting formulas. They are all relative, and I use them so, but if given some combination of IP, H, R, ER, HR, BB, SO, WHIP, ERA, what are the OOTP6 ratings to assign to a given pitcher? Something that could be thrown into a spreadsheet and modified as needed would be greatly appreciated.

[Jayzone]

Originally Posted by Jadetorc

I've seen several threads on the creation, and while this question has been asked before, I haven't seen an answer: What are the formulas for translating PITCHING stats into ratings? Ignoring league totals, ignoring park factors, ignoring defense, L/R splits, endurance and injury. I'm sure they aren't linear, so how they are interrelated is the important issue. I've seen and use calculations for batting formulas. They are all relative, and I use them so, but if given some combination of IP, H, R, ER, HR, BB, SO, WHIP, ERA, what are the OOTP6 ratings to assign to a given pitcher? Something that could be thrown into a spreadsheet and modified as needed would be greatly appreciated.

Well, the problem is that ratings to stats relationships are the product of interactions between ML level batter and pitcher rating distributions with unique league total setting thrown in to boot. As a result, one must use math (regression method) to calculate the 'formula' in each and every different situation/league - as I have done in the utility. So, "great", you said, "just take those formula and use them as if the situation remains constant." Well, problem is that if one wants to implement this in a wide-range scale, then rating distribution will mostly definitely likely to change as the source stats database will likely to have different rating distribution than the league where the formulas are based. Such change in the rating distribution will upset the rating matchup between batters and pitchers and thus render the previously used 'formulas' inaccurate. I would imagine one could possibly use the interation method to do the convert-simulate-convert-simulate... cycles to improve the accuracy of conversion formula in each cycle. However, I have no idea how it might work out in practice.

On the other hand, a single player conversion is certainly plausible in which that rating distribution is not disturbed so that the translation formulas remains accurate. So, alternatively, one could possibly use the gradual method by adding one player at a time and refresh the formula/re-translate the player stats until all players are translated and added.

IMO, accurate wide-scale stats to rating conversion is possible but it require time-consuming processes and result might not even be that accurate.

Er.... I guess I still really haven't answer your questions yet. Well, it's certainly possible to make those dynamically generated formulas from the utility available for users or even provide stats conversion based on those dynamically generated formulas. However, such would require modification of the program and I simply don't have the time for that right now. Sorry.

[Jadetorc]

Well consider this - when the AI imports a set of historical rookies, it will give a certain set of stats a nearly identical set of talent ratings even when imported over and over again. Calculating this talent number without needing to do an import is what I'm looking for. There is a base formula, and it can't be too different to what you're doing, only simpler.
For example, Amos Rusie, with the career stats of :
3769.7ip,3384h,2066r,1287er,76hr,1704bb,1934so,29h bp,153wp,3.07era,1.35whip.

Everytime his rookie year is imported, even into different leagues in different decades, he will import with a talent of 32 32 87 (+/- 1). Does your work happen to have a way for me to generate these results for Pitchers (as I mentioned earlier, i can already do it for batters).

Cheers.