prospect "list of lists"

[bryanbutler]

yeah, i struggled with this, and i've dealt with it different ways in the past. i may regenerate it in this way. what i think i should do is give a blank whatever the max number was in the given ranking +1. so if a given list has 100 entries, then blanks get 101, etc.

i also struggled with excel for about an hour last night trying to make it calculate a reasonable MAD. unfortunately, the way excel deals with blanks when using formulas leaves something to be desired. i finally gave up, and am going to just export the data, calculate it myself in python or perl, and ingest the result into a new column. stoopid excel ><. i'll repost after i've made those two changes.

[rhd]

yeah, i struggled with this, and i've dealt with it different ways in the past. i may regenerate it in this way. what i think i should do is give a blank whatever the max number was in the given ranking +1. so if a given list has 100 entries, then blanks get 101, etc.

i also struggled with excel for about an hour last night trying to make it calculate a reasonable MAD. unfortunately, the way excel deals with blanks when using formulas leaves something to be desired. i finally gave up, and am going to just export the data, calculate it myself in python or perl, and ingest the result into a new column. stoopid excel ><. i'll repost after i've made those two changes.

I think this skews your results. First,, I think a player should be listed on most of the lists before they are included in the composite list. Second, I think a better "fuge value" would be to add a value to the ranking that is about 25% of the total # of slots on the list. E.g., for a top 100 list, if a player is not on the list, give him a rank of 100 + 25 = 125. For a guy not on a top 125 list, give him a rank of 125 + 31.25 = 156. Anyone who ends up w a composite rank of > 100 perhaps shouldnt be listed. You could perhaps have a separate list of guys who made at least 1 list.

[bryanbutler]

I think this skews your results. First,, I think a player should be listed on most of the lists before they are included in the composite list. Second, I think a better "fuge value" would be to add a value to the ranking that is about 25% of the total # of slots on the list. E.g., for a top 100 list, if a player is not on the list, give him a rank of 100 + 25 = 125. For a guy not on a top 125 list, give him a rank of 125 + 31.25 = 156. Anyone who ends up w a composite rank of > 100 perhaps shouldnt be listed. You could perhaps have a separate list of guys who made at least 1 list.

agreed that it skews the results if you have lists of different depth, once you get beyond the number on the list with the least depth +1. some would argue that once you get past 100 it doesn't matter much, but it still seems best to at least *try* to deal with it. i'm thinking about it some more. as you say, it might be best to split it into separate lists (just have different sheets in the workbook).

also fixed a bunch of mistakes (some missing players, some duplicates, some transcription errors, etc.) as i was looking at the lists...

[bryanbutler]

OK, new version. the first sheet has the original, with fixes to mistakes, and median just calculated straight from the cells. the second sheet has the median calculated with blanks replaced by 151 for rhd or SB, and 125 for the others, and no median listed at all for those with median > 125. i could have done some further sorting of those ones, but i think it's reaching the point of diminishing returns...

ftp://ftp.aoc.nrao.edu/pub/staff/bbu...pects new.xlsx

edit to fix bad URL.

[rhd]

I also think it is important to separate the fantasy lists from the "real" lists. I do my own sort of composite ranking, and I do not include the lists geared to fantasy. I also exclude some of the more "questionable" lists. I know that can be arbitrary, but e.g. there is a list that still has Stetson Allie ranked in the top 150. Allie not only is no longer even a pitcher but even when he was pitching he wasnt a top 150 prospect. I dont consider this a reputable list, altho it still is "interesting" to look at. But I exclude it from my composite ranking.

[Moonlight J]

I also think it is important to separate the fantasy lists from the "real" lists. I do my own sort of composite ranking, and I do not include the lists geared to fantasy. I also exclude some of the more "questionable" lists. I know that can be arbitrary, but e.g. there is a list that still has Stetson Allie ranked in the top 150. Allie not only is no longer even a pitcher but even when he was pitching he wasnt a top 150 prospect. I dont consider this a reputable list, altho it still is "interesting" to look at. But I exclude it from my composite ranking.

both the Top 101 real and fantasy lists are coming out at BaseballProspectus next week