Revamping fielding ratings

[DoubleD]

So there's these cool stats by the name of Revized Zone Ratings (RZR) and Out Of Zone (OOZ). RZR is the % of balls fielded within the zone, and OOZ is the amount of balls fielding outside of the zone.

So basically, to make the game more true-to-life, would be good to utilize these stats instead of scouting reports yeah?

Basically I'm thinking RZR = fielding, OOZ = range.

Before I can really starting working on it though, I need to know

What fielding and range rating would you give to an average fielder?

What fielding and range rating would you give to one of the worse fielders (think Miguel Cabrera, Manny Ramirez, Derek Jeter)?

Also just how interested would the main roster guys (Totte and KG) be in this? I know re-editing all of the fielding ratings would be quite a hassel haha.

[kgbaseball]

Also just how interested would the main roster guys (Totte and KG) be in this? I know re-editing all of the fielding ratings would be quite a hassel haha.

Easily. Find me a good basis for fielding and range and I'd redo every player.

[DoubleD]

Haha cool

Can you tell me what you rate the average and bad fielders, as noted in the original post? That's the main thing that'll need to be worked on.

Right now what I'm doing is getting the average stats based off 4 years (rzr/ooz only goes back to 2004). So players around an average rzr/ooz will be given the average ratings ofcoarse...the top guy will be given 99, then I'll figure out how to calculate imbetween average and best.

Same thing for bottom. Obviously you're not going to give the worst guy a 0 fielding rating, so I need to know what the floor would be.

[DoubleD]

btw, you can check out rzr/ooz here: http://www.hardballtimes.com/thtstats/main...g&league_filter[]=2

[kgbaseball]

I don't live and die by fielding stats, but an average fielder is anywhere from 6-9 in fielding, same for average range. Minor leaguers are lower as I like to let progression increase their ratings.

[DoubleD]

ok just a first rough draft of figuring out stats

for 2007 second baseman

http://www.hardballtimes.com/thtstats/main...p;Submit=Submit

for fielding:

15=.880+

14=.872-.879

13=.863-.871

12=.855-.862

11=.847-.854

10=.838-.846

9=.830-.837

8=.815-.829

7=.800-.814

6=.780-.799

5= <.780

(note: average= .830, or 9. Since second base is a high defensive posistion, I think 9 is good for average)

range (IP/OOZ = rate of OOZ plays)

15=34+

14=33

13=32

12=31

11=30

10=29

9=28

8=27

7=26

6=25

5=<25

so for 2b

range   field   name


13	15	Ellis


10	13	Polanco


15	13	Hill


14	13	Phillips


5	12	Giles


9	12	Utley


5	11	Grudzielanek


8	10	Kinsler


13	10	Roberts


11	10	Lopez


7	9	Kent


12	9	Sanchez


15	9	Cano


8	8	Pedroia


9	7	Johnson


6	7	Durham


5	7	Belliard


9	6	Hudson


5	6	Weeks


6	6	Barfield


5	5	Biggio


14	5	Uggla

thoughts?

[kgbaseball]

Looks great; there's a nice range of values (I was worried about everyone being within a 5 number range). Uggla at 14 is sort of surprising; you don't think of him as a defensive whiz if you watch him play.

I'd be very interested in implementing this.

[DoubleD]

Yeah that was the one thing that stuck out. But you also have to look at his fielding. I wanted the floor to be 5, but if it wasn't he'd probably be a 4, maybe even a 3.

I'll do 2006 in a bit, and then average those out and see what they come to, since that'll give higher sample size. Though in 2006 he was also actually an average fielder...man the loss of Perry Hill was huge on Hanley and Uggla.

[DoubleD]

oh man I COMPLETELY f'd up the equation for range originally on my spreadsheet

always double check your work!haha

here's actually what it would be:

range   field   name


8	15	Ellis	Mark


8	13	Polanco	Placido


11	13	Hill	Aaron W


9	13	Phillips	Brandon


5	12	Giles	Marcus


13	12	Utley	Chase


8	11	Grudzielanek	Mark


14	10	Kinsler	Ian M


9	10	Roberts	Brian


9	10	Lopez	Jose C


6	9	Kent	Jeff


5	9	Sanchez	Freddy


10	9	Cano	Robinson


7	8	Pedroia	Dustin L


7	7	Johnson	Kelly A


5	7	Durham	Ray


7	7	Belliard	Ronnie


15	6	Hudson	Orlando


9	6	Weeks	Rickie


11	6	Barfield	Josh L


8	5	Biggio	Craig


11	5	Uggla	Dan C

danny still with a pretty high range here's 2006-2007 combined

range   field   name


5	9	Sanchez	Freddy


7	8	Pedroia	Dustin L


7	7	Johnson	Kelly A


7	7	Belliard	Ronnie


7.5	12.5	Roberts	Brian


8	14	Polanco	Placido


12	14	Carroll	Jamey


13	13	Hill	Aaron W


9	13.5	Ellis	Mark


7	11	Grudzielanek	Mark


7.5	7.5	Biggio	Craig


6	10	Castillo	Luis


10.5	7.5	Uggla	Dan C


9.5	9.5	Kinsler	Ian M


12	10.5	Utley	Chase


10.5	7.5	Barfield	Josh L


6	8	Durham	Ray


9	8	Kennedy	Adam


7	10	Giles	Marcus


9	9	Lopez	Jose C


9	10.5	Phillips	Brandon


8.5	8.5	Cano	Robinson


7	7	Loretta	Mark


9	7	Iguchi	Tadahito


5.5	7.5	Kent	Jeff


15	6	Hudson	Orlando


5	5	Castillo	Jose


7	5	Vidro	Jose


7	5	Cantu	Jorge L

i'm thinking for .5 (and down, since these can be up to 4-year averages), round down, and ofcoarse round up for above.

[kgbaseball]

I could live with those numbers

[kgbaseball]

What are the formulas you're using? I want to do some more testing with this.

[WinnieThaPujols]

OOZ is a counting stat, and so you should probably find the ratio of OOZ per inning or something like that.

Edit: reason being...

Adam Everett - 35 OOZ plays, 535 innings

Alex Gonzalez - 32 OOZ plays, 872 innings

Obviously looking at solely OOZ doesn't do that justice.

Anyway, the biggest problem with using a stat like that (as opposed to a rate stat) is that it's obviously going to be skewed by that player's pitching staff. More strikeout-oriented pitchers = less balls in play = less opportunities for out of zone plays!

With that said, I think this it is better to base something like this on objective stats like these as opposed to just reputation.

Cool project!

[DoubleD]

Yeah, as I said in my post

range (IP/OOZ = rate of OOZ plays)

I also agree that pitching staff and skew things, but it isn't going to have that big of an impact...and in the end, not much else is better haha.

KG:

I don't really have a set formula. Basically I figure out first what I want the average rating to be, then what I want the floor to be.

When I first posted this I set out on averaging the RZR and OOZ rate of inning played each year (instead of just averaging those that qualified). The average rating is then equal to the average RZR/OOZ (So for instance in 2007, 9 is .830 for 2b).

Then basically I just get a rough feel of where the top and bottom of the league were.

For instance, Ellis was at .884 RZR, so I made the max .880+. The two worse 2b (Uggla and Biggio) were at .777 and .776, so I made the floor <.780. Then I just figure out from there

so like....15=.880+, 9=.830, 5=<.780

(.880-.830)/6=.008333..

So then I subtract that from .880 and get .872 for 14, .863 for 13, ect. Same thing for sub 9

(.830-.780)/4=0.0125

So it's kinda of a curve, but i felt that that was the best way to get a wide array of ratings.

here's the averages throughout the years (the formulas have changed which is why the rating needs to be figured out each year and then averaged, rather than adding up the numbers and just doing one rating. This can mostly been seen from LF and RF, the mass difference of average RZR and OOZ)

1b

2007: .741 41 IP

2006: .799 21 IP

2005: .790 22 IP

2004: .753 24 IP

2b

2007: .830 30.2 IP

2006: .820 35.2 IP

2005: .811 29.1 IP

2004: .813 36 IP

3b

2007: .680 25.1 IP

2006: .706 26.1 IP

2005: .735 18 IP

2004: .690 21 IP

ss

2007: .816 22.2 IP

2006: .818 26 IP

2005: .818 22 IP

2004: .823 22.2 IP

LF

2007: .855 26.2 IP

2006: .861 26.1 IP

2005: .633 60 IP

2004: .630 51 IP

CF

2007: .888 22 IP

2006: .894 21.2 IP

2005: .815 22 IP

2004: .796 21.2 IP

RF

2007: .877 27.2 IP

2006: .888 25.2 IP

2005: .648 62 IP

2004: .650 55.1 IP

[DoubleD]

btw I should have a beta of the program out later tonight...it'll only do 2b but I can get feedback on it atleast.

[DoubleD]

ok let me know what you think. Just a small little thing.

Let me know if any errors come up.

BIG *** NOTE:

Make sure you select 2b from the drop down menu (1b is default, so you need to change it). only 2b has been programmed. If any other position is selected, it will not work properly.

[DoubleD]

ok let me know what you think. Just a small little thing.

Let me know if any errors come up.

BIG NOTE:

Make sure you select 2b from the drop down menu (1b is default, so you need to change it). only 2b has been programmed. If any other position is selected, it will not work properly.

other than that, it should be rather strait foward

oh another note

If a player didn't play a year, MAKE SURE that nothing is in that years text boxes. So for instance if a player played in 2005 and 2007 but not 2006, make sure NOTHING is in the text boxes for 2006.

[WinnieThaPujols]

Yeah, as I said in my post

range (IP/OOZ = rate of OOZ plays)

I also agree that pitching staff and skew things, but it isn't going to have that big of an impact...and in the end, not much else is better haha.

Wow, I can't believe I missed that in your post. I feel like an idiot -- my apologies, dude.

[DoubleD]

just a heads up KG, this is going to be on my backburner for a bit (although I basically gave you what you needed to know to do it yourself, if you want to do that rather than wait for the program to make it quicker). I'll hope to have it done within the next couple weeks, not going to be months or w/e...probably work on it during games once the season starts.