A better stat than OPS???

[Ankit]

How about Runs Created per 27 outs (RC27)?

It is how many runs a lineup of 9 of the same player would score per game (27 outs). Using RC27 places more emphasis on creating more runs while using less outs.

The formula is RC27= (Runs Created)/(At bats-hits+caught stealing+sacrifice hits+sacrifice flies+GIDP)*27

The problem with OPS is that it doesn't incorporate Speed, it's a stat that gives more emphasis on power (SLG) and patience (OBP) and only.

I am testing this statistic right now. I have set up a league using Rolen's rosters. After the first season's WS, I will make a second copy of the league and name it Test RC27. The first league will be called Test OPS. I will then reset both leagues back to the start of the season and replay the season. Also, the leagues have Injuries, Financials and Trading disabled.

For the Test OPS league, I will put the player with the highest OPS for their position (CF, LF, RF as considered as on position, OF) and put them on the Philadelphia Phillies in the NL and the Seattle Mariners in the AL.

For the Test RC27 league, I will put the player with the highest RC27 for their position (CF, LF, RF as considered as on position, OF) and put them on the Philadelphia Phillies in the NL and the Seattle Mariners in the AL.

The two teams had the lowest ERAs in their league. Mets had the second best ERA in NL at 3.93, Phillies had 3.91. Mariners' 4.02 was better than the Yankees 4.08.

Here are the teams:

<pre><font size="2">
RC/27
C Paul Loduca, LA C Ivan Rodriguez, TEX
1B Todd Helton, COL 1B Jim Thome, CLE
2B Roberto Alomar, NYN 2B Brett Boone, SEA
3B Scott Rolen, PHI 3B Troy Glaus, ANA
SS Rich Aurilia, SF SS Nomar Garciaparra, BOS
LF J.D. Drew, STL LF Manny Ramirez, BOS
CF Barry Bonds, SF CF Raul Mondesi, TOR
RF Jim Edmonds, STL RF Trot Nixon, BOS

OPS
C Paul Loduca, LA C Ivan Rodriguez, TEX
1B Todd Helton, COL 1B Jim Thome, CLE
2B Roberto Alomar, NYN 2B Brett Boone, SEA
3B Scott Rolen, PHI 3B Troy Glaus, ANA
SS Rich Aurilia, SF SS Nomar Garciaparra, BOS
LF Luis Gonzalez, ARI LF Manny Ramirez, BOS
CF Sammy Sosa, CHN CF Jose Cruz Jr., TOR
RF Chipper Jones, ATL RF Trot Nixon, BOS
</pre></font>

I will put the results in a little while.

How do I make the above lineups show up aligned?

[ March 16, 2002: Message edited by: Ankit ]

[ March 16, 2002: Message edited by: Ankit ]</p>

[JimServo]

You may want to run multiple tests, to cycle out the "random" factors. This is very interesting idea...I wonder what Bill James thinks?

To get the stuff aligned, simply put:
<'pre><'font size="2">, removing the 's, and leave the quotation marks around the number 2. To go back to normal text, type <'/pre><'/font>, againing removing all the 's.

Example, this:
Player AVG HR RBI
Smith .288 2 45
Jones .305 45 125
Baker .210 14 34

Becomes, this:
<pre><font size="2">
Player AVG HR RBI
Smith .288 2 45
Jones .305 45 125
Baker .210 14 34
</pre></font>

[ March 16, 2002: Message edited by: JimServo ]

[ March 16, 2002: Message edited by: JimServo ]</p>

[Ankit]

I have simmed through the season for the OPS league. The Seattle Mariners went 108-54, finishing 10 games ahead of Texas, by whom they would be swept in the Division Series. The team again had the lowest ERA but only the fourth best BA (.287). THey hit 290 HRs, Yankees were second with 246. But they did score fewer runs than them (1106 to 1033). The team had a .859 OPS with .501 for SLG and .358 for OBP. SLG was the highest, OBP was 3rd. The Yankees though had the higher OPS at .866.

In the NL, the story was a lot better. The Phillies went 122-40, 35 games ahead of 2nd place Marlins, 45 ahead of Mets and Braves, 60 ahead of Expos. They hit .297 and had a 3.62 ERA, best by far in both categories. Their 321 HRs were 80 more than 2nd place Milwaukee's total. They scored 1091 runs, 2nd place scored 914! They were tops in 2B but 2nd in BBs. They struck out only 837, fewest in the NL. Seattle struck out the 2nd most in the AL. Their OBP was .365 and their SLG was .536, both the best. OPS was .901 obviously. In the playoffs, Phillies swept the Rockies, beat St. Louis in 7 and then Texas in the Ws in 6.

David Coggin won the Cy because he went 20-3, 2.23 ERA. Luis Gonzalez won the MVP (.292, 65 HRs, 178 RBI, 158 Runs). He led the NL in HR and RBIs. No Seatle player won anything. The player with the highest BA among the OPS teams: Chipper Jones' .340. Seattle was led by Nomar's 128 RBIs while Jim Thome hit 41 HRs. Sosa had a 11.7 RC27, he was second to Giambi's 13.2. Jim Thome's 9.5 led Seattle.

[Ankit]

Bill James created the Runs Created statistic and this is just an extension of it. I will read the New Historical Abstract and see if says anything about it.

[Arbiter]

Sorry to burst your bubble, but RC/27 is essentially proportional to OBP times SLG divided by PA -- which means, essentially, that you're going to get the same results with RC/27 as you will with OPS, at least within the range of normal variation. (One's linear and one's double-log, basically.)

[John C]

I'm thoroughly enjoying the Historical Baseball Abstract now, and one stat has really come to mind:

His win shares.

I think it would be very interesting to see if someone can test this out in a league. I'm thinking about it, provided I can apply the theory.

Much of my ABL history will include a lot of the text that's seen in James' book. It's quite fantastic, and adds many wrinkles to the history that may not have been seen (at least by me).

Peace
John

[Ankit]

[quote]Originally posted by Arbiter:
<strong>Sorry to burst your bubble, but RC/27 is essentially proportional to OBP times SLG divided by PA -- which means, essentially, that you're going to get the same results with RC/27 as you will with OPS, at least within the range of normal variation. (One's linear and one's double-log, basically.)</strong><hr></blockquote>

If it's essentially the same, then why are my NL outfielders completely different in the OPS and RC27 lineups. There are a total of 4 changes among the 16 spots, that 25%, obviously there is a difference. Speed and base running skills do make a difference in the game and its not minimal or insignificant. The lineups above are only for the highest at each position, what if one had to make a decision to choose someone who is middle of the pack OPS in terms of OPS but can steal bases as opposed to some one who is a little better in the OPS department? OPS is good, but RC27 might be a little better and if it is, it will help people make better decisions.

[Arbiter]

[quote]Originally posted by Ankit:
<strong>

If it's essentially the same, then why are my NL outfielders completely different in the OPS and RC27 lineups. There are a total of 4 changes among the 16 spots, that 25%, obviously there is a difference. Speed and base running skills do make a difference in the game and its not minimal or insignificant. The lineups above are only for the highest at each position, what if one had to make a decision to choose someone who is middle of the pack OPS in terms of OPS but can steal bases as opposed to some one who is a little better in the OPS department? OPS is good, but RC27 might be a little better and if it is, it will help people make better decisions.</strong><hr></blockquote>

Well, people have been comparing RC/27 and OPS since at least the late 1980's, so you might find it useful to hit the books before replicating what's been done over and over and over again.

If you play around with the RC/27 formula, you'll notice that RC depends very little on stolen bases and caught stealing, unless a guy both (a) attempts a lot of stolen bases, and (b) has a success percentage that's either unusually high or low. RC/27 does *not* capture any other differences in speed or baserunning ability across players; it is still an open topic just how many runs a guy can add or delete with his non-basestealing baserunning, but most of the data I've seen suggest this would be no more than 5 runs a year.

Finally, OPS isn't even the best OPSlike stat there is; 1.2*OBP + SLG is better. OPS is just more convenient.

I'm not trying to spoil your voyage of discovery, but this is all *old* stuff at this point. What you're doing is essentially trying to pull a signal out of noise where there is no signal.

[Ankit]

You are right about the OPS vs RC, from things I have read, RC using the 1999 is much better than OPS but I don't know which RC formula OOTP uses. However, there are better stats than RC like Extrapolated Runs (XR), and EqA but unfortunately OOTP doesn't make use of EqA, XR is too complicated so its understandable to not include. Now the next question is about setting lineups. Is the traditional method where the best guy hits 3rd, the best way to set lineups? Are there other methods? What about setting them by OPS or RC, highest to lowest, or this way (rank from highest to lowest, either OPS or RC/27):

1
8
2
7
3
6
4
5
Pitcher

Any strategies other people use in OOTP? How successful are they?

[H I Mcdonough]

not to be a jerk, but your outfield should actually go:

LF - Bonds
CF - Edmonds
RF - Drew.

I know that has little or no bearing on this convo, but it just seemed very obvious to me.

H_I