Feedback/ideas On This Method Of Making/updating Rosters ...

[CircleChange11]

Here is an initial idea list that I am sorta compiling to make a 2010 roster set based on projected stats. I would like some feedback on these ideas and suggestions for more efficient ways to get it done.

I missed out on all the great discussions of player ratings in the past, but I have since read all the former threads regrading the topic and picked the ideas that I felt were most doable and would lead to accurate ratings.

Here goes ...

My idea for 2010 MVP Rosters based on “Projected Stats”.

[1] Use 3 different projection systems (MARCEL, CHONE, and ZIPS ... possibly a 4th if I can get either PECOTA or 2009 stats in spreadsheet form), all of which are available in spreadsheet form. Minor league players already have their projections done for the ‘MLB level’, so no need to do MLE’s and things of that sort.

[2] Use 3 different calculation systems (gathered from past posts/threads on player ratings at mvpmods). Most likely using Poo-Diddy's charts, MVPEDit formulas, and PaulW's formulas that were posted in the thread. I would like to use DB Calcs also, but not sure if I have the time to enter all of them into the Calcs. I want to speak to Db about some other alternatives/ideas.

[3] Each projection system will produce 4 different ‘ratings portfolios’ for each player. 4 projection systems x 4 calculation systems = 16 sets of ratings for each player. The spreadsheet does all of the work, otherwise it would be death to the modder.

[4] Take the average of the 16 ratings sets as the ‘final ratings’ (hitters only). This can be done only by the magic of spreadsheet formulas (and

rather quickly too).

[5] Hitters spray charts will be estimated by foxsports “hits charts”, possibly the same thing with G/F. (Unless it is published somewhere that I have not seen yet).

[6] Use the info from fox sports for ONE hot/cold zone for each player. The difference in vR and vL ratings would suffice for “performance splits” ... unles there are published hot/cold zone for vL and vR. I just don't want to guess on it.

[7] Pitchers fielding would be rated as either a 1, 2, or 3, if outstanding fielders getting a 5. This would tend to eliminate the P’s fielding as if they were “hockey goalies”, which hurts the ‘realism’ aspect of MVP. I got this idea from a thread on MVP.

[8] Since projection splits are often just “season total” and not splits. I would likely take the player’s career splits and calculate a ratio to multiply the ratings with to get “splits”. For example if a hitter was better against RHPs than LHPs, I would take their career BA vLHP divided by the career BA vRHP, and multiply that number (decimal) times their contact & power ratings to get the split. Repeat it vice versa to get the vRHP split. Seems this could be done rather quickly … or I might just use asplit estimator” calc to get the ratio (based on career split #'s, if possible). If there's a 'better' or mor 'efficient' way to do this, I'd love to hear it (serioiusly).

[9] Ratings for the other things (Discipline, etc) have been discussed in MVPmods and formulas shared. I’ll use those.

[10] I’m not sure what to do about the swing/take ratings for the various pitches and zones, any ideas? Is that info published somewhere that I don;t know about? I don;t want to guess on this either, but I realize the limitations.

[11] For pitchers, I’ll use the Neyer/James list to check that each pitcher has the right pitches and movement.

[12] I’ll use BM2010 (or 2011) to get each pitchers rating for each pitch they throw and then “mess around” with the MVP combo of control and movement to get the final pitch ratings to match up. (Unless there’s an easier way to do this). This is where I'll spend the bulk of the time, because it's my primary area of interest, and it seems that there are quite a few pitchers in rosters whose ratings have not been adjusted in recent years to accommodate the effect age, injury, or retirment have had on their ratings (or maybe I am missing something). It seems hitters ratings can be calculated in bulk, whereas pitchers rating need to be done "one-by-one" and that is what I (along with help from my son) am prepared to do.

I have not decided whether I would apply this to the UR rosters for public sharing or to KG's rosters for private use (or ask him if it can be released publically). It's just something I'd like to try within the time I have available (limited).

I would likely focus attention on prospects that are in their organization's top 10 (300 total), all of the guys on their team's 40 man roster, and all of the guys listed on the projection lists (many, many minor leaguers), and I would need to create the recent draft picks (which is going to take some work, but it should be 'fun' work.

Thanks for reading, and I really do appreciate all feedback and suggestions as to better ways of doing things. When you guys were doing all this roster work, I was wasting my time trying to get 2k7, 2k8, ad 2k9 to "work right". I don't want to waste more time. In some ways,I'm an 'old pro' at this stuff, and in some ways I'm 'green', so I figured I'd ask the experienced guys for help.

If I think it's 'doable' in a reasonable of time investment to do a "quick and dirty" type of test with 2009 projected stats, I'd like to give it a shot. If not, I'm willing to wait until I have more time.

I would likely "check the accuracy" of the ratings, by using the roster for a 'fantasy draft' and going through the names, by position, and seeing that the names are ranked in what would appear to be a "generally accepted" list .... and/or ... going through each team's roster and observing whether the game places the players in a depth chart as they 'should' (more or less). Then, I would get into simming some seasons and take a look at final stats and make whatever tweaks seem necessary, while allowing for some randomness (I would mostly look at the league leaders and leaguewide averages).

[JoeyMcM93]

That sounds absolutely amazing to me. Major props for wanting to take on such a major project.

For each player's tendencies, I suppose you could get one guy who follows each team to try and take some notes as this season winds down. Otherwise, I don't know how else to do that because it is very subjective.

[CircleChange11]

Fox sports has "hot/cold/neutral" zone date/images, just not for both vL and vR, and I don;t want to guess about it. If I had to guess I'd just leave all zones 'neutral' and let the player's ratings v. the pitcher's ratings determine it.

I just did not know if there was a website/book that contained actual spray chart and g/f ratio's for each player (without being a "scouting service" that require a substancial payment).

[mnkc79]

I have been updating my own dynasty rosters on and off for awhile now, and have found that fangraphs.com (amazing site for a baseball stat fan) has gb/fb/ld ratings for every player. What I do is go into the editor and lets say Mauer has a 37% fb, I will subtract the HR/100 from that # to give him his final fb%. Hope this helps as I am looking forward to trying out your rosters.

[CircleChange11]

I have been updating my own dynasty rosters on and off for awhile now, and have found that fangraphs.com (amazing site for a baseball stat fan) has gb/fb/ld ratings for every player. What I do is go into the editor and lets say Mauer has a 37% fb, I will subtract the HR/100 from that # to give him his final fb%. Hope this helps as I am looking forward to trying out your rosters.

Thanks man.

Another thing I have been doing is keeping a word doument file minimized on my PC, and when I come across suggestions like the one you just made, I copy and paste it into the document, and (when I remember) I add the person's username as reference.

When I get things stuff rolling, I'll compile a list of "Roster making Resources" and post it all in one thread with links to the information, etc.

Edit: Okay, I registered at fangraphs .... umm, yeah .... there's all your pitch data for movement, velocity, control, etc. ' I had a dog and his name was BINGO!' - Ace ventura. Jackpot.