Dog Days Grab Bag

Dave Paisley

It's August. It's hot and muggy, even in Seattle. OK, maybe not Houston hot and muggy, but it's still bad for us. Here goes with a couple of dog days items.

Subject: Bagwell's first run at MVP?

Mac writes:

I'm pretty sure that Bagwell won the MVP in '94. Or don't strike years count?

Then Tom chimes in:

I'm sure someone's already called you on this one, but Bagwell's already got one, for his unreal 1994.

Er, yes, that would be something of an oversight. Looking at the numbers, Bagwell was slugging .750 that year, with a .451 OBP to go with it. Since then he's had a year that matched the OBP, but has never, in any other year before or since, broken .600 slugging. Until now, that is. He also has that OBP around the .450 mark again.

It's odd how my memory of 1994 is affected by the strike. It's much easier to remember the repeat winners than the one-time and never since winners. In the AL, for instance, Frank Thomas winning his second MVP passed relatively quietly because he already had one. Same with the NL Cy Young, where Maddux cruised. It was news to my brain that David Cone won his only Cy Young that year, though. I knew he had a Cy Young award, but if I'd been pressed to guess when he won it. I wouldn't have said 1994.

Perhaps the most heinous loss from that season is Montreal's run at the division title. Technically they won it, but without the usual conclusion to the season (playoffs), it doesn't really stick. Is their division "title" tainted? If it is, are the individual awards tainted as well?

I'd like to believe not, but my heart says that the awards that year, while they actually happened, just don't mean as much. Anyway, sorry Jeff. Even if it wasn't obtained by the optimum route, it still counts.

More Stat Stuff

After checking out the stolen base side of Runs Created last week, I was motivated to take a closer look at Bill James' Runs Created stat (usually abbreviated as just RC) in some more detail. The simplified version of the RC calculation that leaves out sacrifices and sac flies and so on is made up as so:

A = H + BB - CS

B = TB + .52 * SB + .26 * BB

C = AB + BB

RC = A * B / C

Where:

H = Hits

BB = Walks

CS = Caught Stealing

TB = Total Bases

SB = Stolen Bases

AB = At Bats

Many RC proponents argue that simpler measures, such as OPS (On base Plus Slugging) are much inferior to Runs Created, especially in the sense that OPS undervalues on base percentage (OBP) and overvalues slugging (SLG). I decided to see for myself just how the two compare in a neutral environment with some hypothetical players.

First, here's a table where I steadily increase OPS, while keeping the ratios of batting average, on base percentage and slugging constant. The OPS values range from .600 (slightly above Rey Ordonez) to 1.150 (Mark McGwire in a good year.) I also adjusted each line to 600 plate appearances (PA)

OPS	OBP	SLG	H	BB	TB	RC	R/OPS	Diff
.600	.240	.360	108	36	203	52	47	5
.650	.260	.390	117	39	219	60	58	2
.700	.280	.420	125	43	234	69	69	0
.750	.300	.450	133	47	249	79	81	-1
.800	.320	.480	140	52	263	89	92	-3
.850	.340	.510	148	56	277	100	103	-3
.900	.360	.540	155	61	291	112	114	-3
.950	.380	.570	162	66	305	123	126	-2
1.000	.400	.600	169	71	317	136	137	-1
1.050	.420	.630	176	76	330	148	148	0
1.100	.440	.660	183	81	342	161	159	2
1.150	.460	.690	189	87	354	175	171	4

You can easily see that RC increases steadily from a paltry 52 runs, through to a whopping 175 runs. So it's interesting to note that doubling OPS results in almost four times the runs. Remember that the next time you wonder why your team is paying much more for the .900 OPS guy than the .800 OPS guy. Or why you should be upset that they're paying the .800 OPS guy the same money they could have got the .900 OPS guy for.

However, I've also shown how runs correlated with OPS match up with the RC formula. It's in the next column, labeled R/OPS. It's a little tricky because these are hypothetical players, and RC is a hypothetical stat itself. Nevertheless, the correlation of runs with OPS to RC is extremely good. There's a slight non-linearity, such that OPS underestimates at the extremes and overestimates in the middle, but the difference is, at worst, about 3%. Nothing to worry about in the grand scheme of things.

So, given that, which would you rather calculate - OPS or RC? Gee, add two numbers together or retake high school algebra… Tough choice.

Next time, I'll examine the relative importance of OBP and SLG for a fixed value of OPS to see if there really is much of a difference exactly how a player gets his OPS.

about the author

Dave Paisley just loves this stuff. To keep him busy this summer, please send him all the Excel spreadsheets you can at drdjp@jetcity.com.