An Easy and Quick Fix for Defense in X?

[knockahoma]

I've been perusing baseball-reference.com, lately.

They have a categoy called RTOT/YR. Somebody else's post turned me onto it (Pstrickert, I think?)

This is the number of runs saved by a player's defense above, or below, an average player. So, you get numbers like "-7.6 runs saved below average player", or "+27.5 runs saved above average player."

Currently, OOTP calculates a player's defense by range factor and field percentage (unless I'm badly mistaken).

Rather than use range factor, I wonder if RTOT/YR would be the better choice? It takes statistical snap shots from several more angles. For example, RTOT/YR includes:

* RTZ. Field plays made, not including double plays, or outfield kills.

RTZrd Plays made above, or below, the average player on the road.

RTZhm. Plays made above, or below, the average player at home.

RDP number of double plays turned and opportunities given

RCatch. No explanation on that one that I saw.

RTOT/YR bases its calculations on 1200 innings played in a year, or approximately 135 games.

So, one can infer range out of this, and ability to turn double plays, etc. Now, you can get some goofy numbers for a player who just stepped in for a few games, but you get unrealistic numbers from just range factor in that situation, too.

As I was perusing the 74 season, I saw Mark Belanger with great runs saved above the average player numbers (from +17 to +34). Good hitting shortstops with reputations for average or poor play (Toby Harrah) might get a -8, or -10 score. And so on. What's interesting is that these numbers correlated extremely well with the Strat-o-matic ratings I remember for various players (Strat has always put a great deal of research into its defense ratings, using numbers and scouts, etc.).

Rather than OOTP creating its own defensive database of all players, could we get immediate and satisfactory results by switching from range factor (or maybe including that) to RTOT/YR, with its various sub-categories already included?

I don't know if Lahman includes this stat, but it would be a simple thing to get the numbers. They're already at baseball-reference.com.

Then, get OOTP to evaluate and rate players on RTOT/YR (in addition to range factor, maybe?) and field percentage. Not only would we get better defensive ratings, but OOTP can then better evaluate the defensive contribution of a player when making trades.

All without having to create a specialized defensive database. Unless...OOTP already uses this information. If so, it's back to the editing page.

[Afino]

I also noticed this in baseballreference as well.

Ironically, while also looking at Mark Belanger for an entirely different reason.

I agree that it should be in OOTP in some way. For a basis of fielding ratings, as a stat, however.

[knockahoma]

Yeah,

I think it might do well as part of the big three. Range factor, field percentage and RTOT/YR, which is just a conglomeration of other stats.

What got me excited was how RTOT/YR could help us telescope in on defensive skills AND gives the AI more info for trades, etc. Now, the AI would consider the issue of runs saved as well as runs produced. That might even help with lineups, and the AI's lack of appreciation for fielding at SS--- in OOTP 9, anyway. I'm sorry. I just can't see 36 year old Brooks Robinson getting a start at SS every 10 games, as happened in one of my 74 sims.

And satisfaction might come quicly. Adapting OOTP to reading this stat must be ton easier than creating one's own database.

[StyxNCa]

Something definitely needs to be changed. Even using a players att/9 innings rather than att/game would help some.

[pstrickert]

OOTP has the following categories for infielders:

Infield Range (= Rtz)
Infield Arm (= ???)
Turn DP (= Rdp)
Infield Error (= FldPct)

Rtot is an overall rating (I think)

Correct me if I'm wrong, but I think RF/9 only considers the number of plays made. Rtz takes into consideration the percentage of possible plays made.

[StyxNCa]

Originally Posted by pstrickert

OOTP has the following categories for infielders:

Infield Range (= Rtz)
Infield Arm (= ???)
Turn DP (= Rdp)
Infield Error (= FldPct)

Rtot is an overall rating (I think)

Correct me if I'm wrong, but I think RF/9 only considers the number of plays made. Rtz takes into consideration the percentage of possible plays made.

You're correct but if it does use Rtz the ratings make even less sense. The only thing I can think of, for earlier seasons, is that it's basing the rating on the fielding stats of today and not basing them on the actual league year that is being played.

[pstrickert]

OOTP uses RF/9 for Infield Range, FldPct for Infield Error. I placed in parenthesis the baseball-reference.com stats which could, instead, be used for the OOTP categories.

[krelian]

It would be great if OOTP allowed you to add custom stats...

[knockahoma]

Boy, RTOT and RTOT/YR are fast becoming some of my favorite stats. The certainly reflect a lot of reputation from the time. In 1974, for example, the team RTOT scores looked like this in the final standings.

BAL 43.6 (above the average player runs saved number)
NY 21.7
BOS 46.1
CLE 37
MIL 18
DET -26

Baltimore's strong defense saved more runs than any other team in the AL, save Boston. Baltimore's defensive stars, according to this stat, were Belanger, +17.8, Paul Blair, +14.6, Brooks Robinson, +13.0, Grich 8.2. Earl Williams was an unspectacular -4.2

Boston's defensive stars were Dwight Evans +27, Burleson +12 (he took over for Mario Guerrero that season at -2.6. Juan Beniquez CF +11.6 and Rick Miller, +15.8.

What's been really interesting, and I think complimentary of baseball people's perceptions in that era, is that players being moved because of defensive liability correlates very well so far with this stat.

For example: Bobby Grich was tried at SS in 77. He scored about a flat 0 in the category. He was moved away from shortstop. Boston's Rico Petrocelli used to be a good scoring SS in RTOT. In 1970 he seemed to lose range and scored a +2.0. He was moved to 3B the next year.

Earl Weaver believed that his gold glove 2B Davey Johnson had lost range and decided to trade him to Atlanta, in part, for that reason. Johnson scored a +13 in 1972. But he fell to a -8 in 1973 in Atlanta, and soon after became primarily a 1b for Atlanta and Philly.

Toby Harrah was a monster hitting shortstop in 74 with 21 homers. But the rep was that Texas gave up too much on defense. Harrah played SS for texas through the mid 70s with scores of -8 and -10. Finally, Texas moved him to 3B and got Bert Campaneris at SS.

The AL West in 74 ended like this:

Oakland 38.8

Texas -28.8 (5gb)

Min -20

chic -39.3

kc -25

cal -25


So far, this stat makes old school baseball people look good in many instances.

[StyxNCa]

Originally Posted by pstrickert

OOTP uses RF/9 for Infield Range, FldPct for Infield Error. I placed in parenthesis the baseball-reference.com stats which could, instead, be used for the OOTP categories.

Ah ok. My bad.

[actionjackson]

Originally Posted by knockahoma

...Boston's defensive stars were Dwight Evans +27...

The AL West in 74 ended like this:

Oakland 38.8

Texas -28.8 (5gb)

Min -20

chic -39.3

kc -25

cal -25


So far, this stat makes old school baseball people look good in many instances.

I was only 4 for most of the 1974 season and didn't get into baseball much until the Jays arrived three years later. Then it was: hockey? What hockey? Well, not quite. That came later during the explosion of 24-hour sports. The media up here's always been about 90% hockey, or so it seems to the hockey-fatigued, hanging in to catch a brief, fleeting glimpse of baseball/football/basketball/tennis/soccer/hot dog eating contests/etc...and now back to hockey! Groan! Thank goodness for the interwebs.

Good to see Dewey get his due here, and Grich for that matter. That AL West looks pathetic defensively, save of course Oakland. A stat that brings together stats geeks and old school baseball people? Awesome. Now maybe all the territorial p*** fights can simmer down a bit more and we can get back to enjoying baseball instead of bickering.

[knockahoma]

You know, there's a lot of talk about stats that are predictors. I've just been delving into RTOT today. The years 74-76 and 83-87 are my favorites. I've played thousands and thousands of games in those periods with various baseball applications (APBA, STRAT, MOGUL, OOTP). Plus, I was an avid baseball fan in that era. So, I kind of know those rosters inside out.

Putting aside for a moment whether managers had it right on philosophy back in the 70s, the predictive ability of this stat regarding trades has been fascinating so far-- and I might add, a real testament to STRAT's defense ratings, an area of immense pride for the company for years.

The stat appears early on to be a great predictor of moves based on a particular philosophy.

AL EAST 74

Oakland 38.8

Texas -28.8 (5gb)

Min -20

chic -39.3

kc -25

cal -25

Texas people apparently perceived this "run saved" issue through watching games and other stats. In the next few years, they found a good CF (Beniquez, an eventual gold glover) and moved Harrah to 3B, picked up Campy at SS, a year after trading their other shortstop prospect, Roy Smalley, to Minnesota for pitching help.

Meanwhile, Minnesotat took 2b Carew (-6.5 in 74 and 4.1 in 75) and moved him to 1B, where he scored a +9. In 1975, Danny Thompson scored horribly at SS (-14). So, Minnesota traded for Texas prospect Roy Smalley (+5.2 in 76) and set Bob Randall at 2b (-1.8, +2.9 and +14.1 in his first 3 seasons with Minn).

So, looking at the infield in 74 for Minn, 1B Kusiak, 2B, Carew, SS Thompson, Minnesota scored about a -5.

Minn (sensing the difference between Baltimore and Oakland in this area?) tried to improve its infield defense. In 76, with Carew at 1B (+8.9), Randall at 2b (-1.8) and Smalley at SS (5.2), Minnesota's differential in the category was about +12.7 runs saved above average player, just with those 3 positions changes.

When you look at the leaders in that category in 74 (Bost 46, Balt 42, Oak, 38), that's a significant percentage.


Over in the east in 74,

BAL 43.6 (above the average player runs saved number)
NY 21.7
BOS 46.1
CLE 37
MIL 18
DET -26

the Yankees seemed to see their own mediocre score in a category that had yet to be devised. The Orioles infield of

2b: Grich (6.8)
SS Belanger ( +17.9)
3b Robinson (13.0)

finished with +37.7 runs saved over the average players per year.

Meanwhile, the Yanks finished with

2b Sandy Alomar (4.8),
SS Jim Mason (-5)
and Graig Nettles (13.0)

for a total of +13.2.

The Yanks started looking for a double-play combo to beef up a score they didn't know existed.

They traded for 2B Willie Randolph and got SS Bucky Dent. In 1977, when the Yankees newest dynasty began after a disappointing 76 series showing, the infield score looked like this:

Randolph +11.6
Dent 10.7
Nettles 17.5

+39.8 runs above average player saved... from those three positions alone.


Two observations. These numbers can get pretty wild looking with just a smaller playing times. But then, most numbers can. The general idea still holds true if you look at career numbers in this category. A part of me hates to do that, since it flattens out some great players who get old, failing to reflect some great seasons for what they were.

[pstrickert]

Originally Posted by knockahoma

Putting aside for a moment whether managers had it right on philosophy back in the 70s, the predictive ability of this stat regarding trades has been fascinating so far-- and I might add, a real testament to STRAT's defense ratings, an area of immense pride for the company for years.

By contrast, fielding is a GLARING problem in OOTP. Plus, if anything is done, it won't be until next year.

[pstrickert]

knockahoma, can you do the same analysis for the NL East and NL West? I suppose I could do it myself, but I'm about to leave for a couple of hours. Thanks! I thoroughly enjoy reading your posts.

[actionjackson]

Originally Posted by pstrickert

knockahoma, can you do the same analysis for the nl east and nl west? I suppose i could do it myself, but i'm about to leave for a couple of hours. Thanks! i thoroughly enjoy reading your posts.

+1

[Vogon Poet]

Originally Posted by pstrickert

OOTP has the following categories for infielders:

Infield Range (= Rtz)
Infield Arm (= ???)
Turn DP (= Rdp)
Infield Error (= FldPct)

Rtot is an overall rating (I think)

Correct me if I'm wrong, but I think RF/9 only considers the number of plays made. Rtz takes into consideration the percentage of possible plays made.

I'd love to see OOTP switch to the Fans Scouting Report position-neutral model: The 2008 Scouting Report by the Fans for the Fans

Instincts
First Step
Speed
Hands
Release/Footwork
Arm strength
Arm accuracy

It would be more descriptive, and make position changes so much easier to evaluate.

[knockahoma]

knockahoma, can you do the same analysis for the NL East and NL West?

Being a Braves fan in that era, no, I can't. I won't go back to that.... JK. Sure! I was going that way, anyway. And thanks for the nice words, guys. That was quite a shock to discover someone else found this interesting.

[knockahoma]

Continuing the look at Teams through the prism of the TOT/LG stat:

1974 NL West

Dodgers 3.4 TOT/YR
Reds 5.4
Braves 6.0
Astros 0.6
Giants - 0.3
Padres -8.4

The first thing that strikes me is that while, pitching and defense were long aspects of pride for LA, the Reds looked better in both TOT/YR and field percentage. In fact, LA finished a point under league fielding average. The Reds finished just 4 games back, the only real threat to LA.

In actuality, the reps of defensive players remain intact. It was the horrific performances off the bench by Lee Lacy, Von Joshua, Ken McMullen, that brought down LA's score. Jimmy Wynn finished with a nice +12 in CF.

Meanwhile, the Braves had discovered in 1973 that, despite Davey Johnson's bizarre and once in a life-time output of 40 homers, Earl Johnson was right about the second baseman's deteriorating range. He finished with a -8. His double play partner, Marty Perez, had upped his score in 72 (a dreadful -13) to league average. But, together, the Johnson/Perez combo had given up -8 runs below league average. In 74, they moved the infield around, depending on if they faced a righty, or a lefty. Johnson split time between 1B (+8) and 2B (-10). Perez played almost exclusively at 2B where his scores vastly improved (+11, .985%) Robinson moved in at SS and played at league average (+1). But the Braves (based on 135 games) improved greatly at these positions over 73's total. +7 at 1B, +19 at 2B and +1 at SS, for a total of 27 runs saved above average.


The Astros:

Roger Metzger is the surprise to me. In some baseball sims, Metzger was awarded top range. But, by his life-time TOT/YR is only -4. He was a very steady shortstop, though. Once again, I wonder if some gifted athletes at the position flashed with dives, but were often out of position? He had just won the gold glove in 73 with a -4 TOT/YR score. But, he made only 12 errors the entire year.

Doug Rader was the same type fielder. Five gold gloves, only 16 errors in 74, compared with Schmidt who ended Rader's GG dynasty with error totals often in the 20s.

Catcher/1B Cliff Johnson's score makes the traditionalists happy (-19), cancelling out his 10 homers in part-time duty. Lifetime, he was -14 at 1B, -15 at catcher and a hideous -22 in the OF. The Astros knew his 20 homers in 75 probably didn't cover the statistical and psychological damage caused by Johnson's defense. The bat, in their view, wasn't worth the cost.

The Astros sent Johnson off to DH duties in the AL for middle infielder Mike Fischlin.

Houston's other defensive problem was LF Bob Watson. A fine fielding percentage didn't hide the flaw (a consistent -15 over the past 3 years). It's interesting to note that Houston had here the controversial situation. Throw your bats out there and eat the defensive cost, with 1B Lee May, LF Bob Watson and C Cliff Johnson, or shore up your defense. They chose the latter:

Lee May was shipped to Baltimore. Watson moved to 1B (-8) The Astros traded for speedster Jose Cruz, a disappointment in St. Louis, but a solid glove (+9) in 75. Cesar Cedeno was gold glove material and Greg Gross offered a line drive bat with a -3 rating. In the infield, they moved prospect Rob Andrews (-8) into 2B, down a total of 14 points from Tommy Helms the previous year.

The result in runs saved? Totaled, a drop to -1. Watson suffered at 1B (-8), Cedeno had an off year (-4) and Andrews, categorically speaking, was a disaster from Helms, a former multiple gold glover and .279 hitter the year before (Helms was still with Houston. An injury situation?).

The Astros finished last in 75 with a miserable 64 wins.

But in 76, Houston made gains with its new lineup, finishing in 3rd with an 80-83 record. They finished 2nd in the NL west in runs scored, but far behind the Big Red Machine in runs (3.86 to 5.29). On the Astros staff were solid pitchers, such as JR Richard, Larry Dierker, Joquin Andujar, Ken Forsch and Joe Niekro.

It took until 1979 for Houston to reach prominence in TOT/YR. They finished +3.1, 2nd in the category in the NL. Their record was 89-73, 1 and 1/2 game back of the Reds in the West.

In a 2006 article, Bill James alluded to the fact that saber-theories were missing something with fielding. The swings were too wide. The obvious inference would be that balls on play is controlled much more by pitchers, or luck than thought. But something appears to be missing when a Cesar Cedeno swings from a +8 to a -4 in one year.

I see those kinds of inexplicable swings looking over these seasons, though the numbers do correlate strongly with reputation of players at that time. And the category continues to be a good indicator of how the defense will be shuffled by teams during the off-seasons.

More to come, if anybody remains interested in this subject.

[pstrickert]

I hope it's OK to post part of Garlon's proposal for historicals here. (Markus incorporated other parts of his proposal in OOTPX, but -- as far as I know -- not this part.) I think it addresses the problem we've discussed in this thread.

8. Overall Team Defense Impact on opposing Batter BABIP rating

NOTE: If we include this in the game we must be very careful to do it correctly. It should be the very last thing we add for historicals if we do it at all. The rest of the fielding improvements will work just fine without this last Modifier.

For historicals we can put one more finishing touch on the fielding dynamics. One thing that somewhat concerns me is how a team's overall defense affects turning a ball in play into an out. Certainly teams with above average defenders at all positions should have a better DER (Defensive Efficiency Record) than a team with all average defenders. So using MLB data from 1996-2006 I compiled data on plays made by each position. What we can do is have OOTP look at the sum of a team's total defensive ability that is on the field and have it affect the likelihood of a hit on a BABIP...this would essentially be modifying a hitter's BABIP rating. Pitchers who have superior defenses behind them compared to the rest of the league will end up with BABIP against that is slightly below league average and vice-versa. This system completely coincides with how I designed the fielding ratings system and the modifiers.

What we can have OOTP do is add up each player's Modified Range Factor (PO+A+DP)/G for the current position they are being used at on the field. So OOTP would look at who is curently playing 1B and find his modified range factor, then look at the 2bmen and find his modified range factor and so on across all 9 positions and add these up. Remember, my Overall Positional Calculation is a direct translation of Modified Range Factor so Modified Range Factor could be backward calculated from the Overall positional Ratings too. We add up the totals for all 9 positions and then divide by 37.2.

Ok, so let's say a team is fielding a defense whose sum Modified Range Factor per Game is a 38.3. We'd take 38.3/37.2 = 1.029 This means that this particular team is 2.9% more likely to turn a ball into an out than an average team. So if the league average BABIP is .300, this team would have a BABIP against of .291 assuming of course that the average Modified Range Factor for all the other teams in an average game over the course of a season is actually 37.2.

If in this example, however, all the other teams in the league were also fielding teams with a 38.3 rating in their average games, then this team in question would be league average in the end. The LTM's for Batting Average and such hold things in check.

The only thing that OOTP would need to do is check the current on-field Defensive Efficiency Rating and use that to multiply a batter's BABIP rating in each AB of the game so that we can account for defensive changes.

Defensive Efficiency Rating formula for OOTP:

(HBA)/(TMR/37.2) = Adjusted BABIP for a hitter

HBA = Hitter's BABIP
TMR = Team's Sum Modified Range Factor
37.2 = Baseline League Expected Sum Modified Range Factor for Teams

Ok, so let's say a hitter has a .280 BABIP ability. And let's say that a team's sum modified range factor is a 38.3.

.280/(38.3/37.2) = .272

This batter will hit .272 on Balls In Play against this team instead of .280.

The formula could also be written as (HBA * 37.2)/TMR

This allows for the overall quality of a team's defense to be factored into the outcomes while at the same time allowing the distribution of Positional Ratings of the players on the field to determine who actually makes a given play after an out is calculated. Team's with higher sum DER will help their pitchers turn more balls in play into outs. With this system it wll now be possible to field quality teams based around pitching and defense in historicals with only 1 extra multiplication factor added into the equation of BABIP for each hitter. This is DIPS at it's best in historicals since pitchers don't control BABIP - the hitters and the team defense do.

If a player is playing out-of-position and therefore does not have a Positional Rating to translate back into a Modified Range Factor, assume these Modified Range Factors for the out-of-position:

P: 1.36
C: 5.43
1B: 7.60
2B: 4.21
3B: 2.14
SS: 3.86
LF: 1.51
CF: 1.95
RF: 1.60

A team with all of their players out-of-position would end up increasing opponent batting average by 25%.