Roster Creators.....Arm and Range

[LilRedMachine]

How can these be derived from stats? Range totally mistifies me. All I can tell for range is what I hear/read about a guy. For instance, Ranier Olmedo (Reds SS prospect, watch for him in a couple of years) is being talked up as the best SS in the organization by alot of folks in Cincy, as far as glove and playmaking ability. What stats, other than Fielding %, back this up. Is range a product of actually watching a guy play and saying 'Wow he really made a tough play look easy'?

Arm for outfielders gives me the same headache. I suppose you could base it on Assists, but, if a guy has a great arm, he may not get the opportunity to get alot of assists. For example, not many guys are gonna run on Raul Mondesi(good arm), but put Dieon Sanders out there and watch 'em go.

Arm for catchers does seem a little more straight forward. I think you could base it almost soley on % of runners thrown out vs. league average. I'd be glad to hear others thoughts here though.

[Erich Ingram]

I just go by what I can read up on a player. If I read a scouting report that says has a cannon no one runs on him then he gets a higher mark. When I read couldn't through out a turtle then they get a lower grade. A lot of it hinges on watching the players as well. I didn't not change a lot of arm ratings from what the computer generated. It is something that I want to go through and look at still. Guys like Kotsay I don't think have the right arm strength right now as he should be an A and I don't think he is.

[locust9]

range can be derived a variety of ways. Bill James' range factor formula is probably the most accepted.

Its formula is (Putouts + Assists)/Games or divided by 9 innings. I'd like to see marcus put this in here - it really helps when you're eyeballing a fielder (and potentially, it could be used as another qualifying mark for the HOF). As it is right now, fielders get short shrift in Marcus' game. Ozzie Smith and players like him would never make it in.

/l

[CBL-Commish]

Problem is that range factor is hugely dependent on other factors besides a player's real range. Handedness of the pitching staff causes more balls to be hit to one side of the field or the other. Groundball/flyball ratio of the staff changes infield and outfield chances. A gap in the quality of the outfield compared to the infield (or vice versa) can subtly change the number of putouts each gets.

For first basemen, 90% or more of all putouts are just catching a throw from someone else, standing on the bag. Assists are dependent on whether or not the guy likes to take it to the bag himself, or just wait for the pitcher to cover. For catchers, most putouts are just catching the third strike. For third basemen, putouts are highly dependent on how much foul ground there is in his home park.

There are ways to compensate for all of these biases - the newest work from Baseball Prospectus and Bill James each do a good job. But the methods are complex and time consuming. Subjective analysis may be the easiest thing to do now, but in a month or so James' new Win Shares book comes out, supposedly with defensive analysis of every player in history. That could be very valuable in rating real range.

[GForce22]

Reading scouting reports is the best way to go (obviously, for minor leaguers, the only way to go, since fielding data is so hard to come by). Assists, as was mentioned, are not a good indicator. Find as many scouting report sources as you can and that's usually the best method.

GH

[locust9]

agreed. However, park factors and all else aside, range factor is still a pretty useful stat, certainly as useful as some other stuff I've seen included here.

I had no idea james had a new book coming out, that's great. Any idea when? I've been going nuts trying to find any of the abstracts I'm missing on antiquarian bookseller lists: probably because freaks like us are busy hoarding them.

I really wish marcus would resurrect runs created for career stats. I find it a much more readable stat than that wacky ops one, and rc per game is good to know.

/l

[CMH]

i would visit mlb.com, somewhere in there they probably show how they come up with a players range. it was mentioned in a post above, but putouts, assists, errors, etc. all play a role. i know that once a number is found it is compared against the league average range. in the result, a range value is found. i dont know how though, but its all or some of the things i just mentioned.

[CBL-Commish]

[quote]Originally posted by locust9:
<strong>I had no idea james had a new book coming out, that's great. Any idea when? I've been going nuts trying to find any of the abstracts I'm missing on antiquarian bookseller lists: probably because freaks like us are busy hoarding them.
/l</strong><hr></blockquote>

Apparently sometime this month.

Here it is:
<a href="http://www.stats.com/store/store.asp?page=wins" target="_blank">http://www.stats.com/store/store.asp?page=wins</A>

[LilRedMachine]

OK, this morning while trolling through stats on ESPN.com I found something called Zone Rating.

Zone rating. The percentage of balls fielded by a player in his typical defensive "zone," as measured by STATS, Inc.

Is this basically ball fielded in 'zone'/balls hit to 'zone'? Maybe this is a good base for OOTP4 ranges.

[CBL-Commish]

[quote]Originally posted by LilRedMachine:
<strong>OK, this morning while trolling through stats on ESPN.com I found something called Zone Rating.

Zone rating. The percentage of balls fielded by a player in his typical defensive "zone," as measured by STATS, Inc.

Is this basically ball fielded in 'zone'/balls hit to 'zone'? Maybe this is a good base for OOTP4 ranges.</strong><hr></blockquote>

Yea, zone rating is a pretty good indicator of range, but subject to biases like anything else. Stats has scorers watching every game and making subjective calls on what goes to each zone, I think. They've had some discussions and changes in methodology relating to line drives, fly balls and how each is rated or taken into account. But anyway, ZR is probably as good as anything currently available for extracting OOTP range info.