Tuesday, April 21, 2009

Mythbusters: Using UPL VORP

If you read through that last discussion regarding the attempted Mauer/Lester/Dukes for Berkman/Holliday trade, you'll see a lot of statements that are worth looking at.

One excerpt:

"Berkman career is solid, but year to year he is a wild card trust me. I have had him in other keeper leagues before.

Is berkman better than MAUER considering position scarcity. I doubt it."

So, we should really look at this. We've talked about the concept of Value of Replacement Player (VORP) in a UPL context. Basically, one way to evaluate a player is to see how much of an additional benefit it is to have a player, relative to what's available in free agency. Of course there are issues with this concept, particularly with regard to who's available - people will often keep bad players on their roster for too long. And I'm not going back to see who was actually available in 2007 and 2008, since that info isn't readily available. But it's a nice start point.

Let's look at some numbers at C, specifically the best stats from current free agents in 2007 and 2008 (which is what Yahoo! has available).

Ivan Rodriguez (2007 - 502AB): .294/.420, 50R, 11HR, 62 RBI, 2SB
Ramon Hernandez (2008 - 463AB): .308/.406, 49R, 15 HR, 65 RBI, 0SB

Certainly, nothing overwhelming.

Now, let's look at Joe Mauer in that same time period:

Mauer (2007 - 406 AB): .382/.426, 62R, 7HR, 60RBI, 7SB
Mauer (2007 - 536 AB): .413/.451, 98R, 9HR, 85RBI, 1SB

Clearly, Mauer's better than a replacement player, which was never in doubt. But how much better?

Typical Replacement Season (482 AB): .301/.413, 50R, 13HR, 63RBI, 1SB
Typical Mauer Season (471AB): .397/.439, 80R, 8HR, 73RBI, 4SB

Difference? Basically, having Joe Mauer, compared to a free agent? About .008 in OBP, .002 in SLG, 30R, -5 HR, 5RBI, and 3 SB.

So, if what Phatsnapper's saying is true, that Mauer is better than Berkman, considering position, then we'd have a lot of people screwing up their drafts, since Berkman comes off the board around 18, whereas Mauer comes off the board at 55. The most likely explanation is that people don't evaluate position scarcity accurately. This could be entirely true.

Now, let's look at Lance Berkman and a replacement 1B.

Berkman (2007 - 561AB): .386/.510, 95R, 34HR, 102RBI, 7SB
Berkman (2008 - 554AB): .420/.567, 114R, 29HR, 106RBI, 18SB
Typical Berkman (557AB): .403/.538, 105R, 33HR, 104RBI, 13SB

Oddly enough, in 2007 and 2008, the best replacement looked to be Ryan Garko:

Garko (2008 - 495AB): .346/.404, 61R, 14HR, 90RBI, 0SB
Garko (2007 - 484AB): .359/.483, 62R, 21HR, 61RBI, 0SB
Typical Replacement (490AB): .351/.444, 62R, 18HR, 76RBI, 0SB

The difference between Berkman and Garko is about this: .004 OBP, .008 SLG, 43R, 15HR, 28RBI, and 13SB.

The difference between differences, for a given season?

Mauer Benefit (2007-08): .008 OBP, .002 SLG, 30R, -5HR, 5RBI, 3SB
Berkman Benefit (2007-08): .004 OBP, .008 SLG, 43R, 15HR, 28 RBI, 13 SB
Benefit of Mauer - Berkman: .004 OBP, -.006 SLG, -13R, -20 HR, -23 RBI, -10 SB.

So, is Mauer, compared to a replacement catcher, better than Berkman, compared to a replacement 1B? I'd guess no. What does this suggest? Well, based on this small sample size, the wisdom of the crowds, and the wisdom gleaned from quick research, appears to be better than the wisdom of Rup's guesswork.

But what about down the road? How will this trend move? If you give both players the benefit of the doubt, and say that they'll be relatively healthy in their careers, you'll see Berkman start to regress, and Mauer improve a little bit (or maybe just stay at his 2008 levels). I'd guess that Berkman falls to Mauer's level (which should be entering the plateau stage) in about 4 seasons, and when Berkman is looking at retirement (after 5 or 6 seasons), you'll still have Mauer at a high level for 2-3 more seasons, before his starts his decline.

However, if you look at prior health, oddly enough, from 2005-08, Mauer has had 2237 AB + BB, while Berkman has had 2501 AB+BB. What this suggests is that Berkman has been healthier than Mauer (and is right now, as well). I don't think that this changes Berkman's career trajectory, however, Mauer may be suspect to a faster decline. Also note that he's good for about 15-20 games at DH, so his numbers are based on being on an AL club. If he's on an NL club, figure that he'll be catching 130 games, so you have to take his projections down accordingly. And if Mauer moves positions, then his value diminishes immensely, and this whole discussion is moot. On the other hand, if Berkman moves to the AL, he'll have the luxury of playing DH.

So a very appropriate question is between Mauer and Berkman, who's the larger injury risk? Generally, you'd say that 33 year old is more at risk than the 26 year old. But given their individual histories and the position that they play, it's awfully close, isn't it? But for the next 4 years, it's very likely that Berkman will be the more valuable player, even after you adjust for position. After that, it depends a lot on Mauer's health.

-Chairman (aka O.N. Thugs)

3 comments:

clauff said...

i wonder how your analysis factors in when you have a backup like Weiters on your bench, waiting to be called up. I guess, VORP analysis holds true until Weiters plays.

This is probably why he can irresponsibly claim that Mauer and Berkman are equivalent players.

clauff said...

i wonder how your analysis factors in when you have a backup like Weiters on your bench, waiting to be called up. I guess, VORP analysis holds true until Weiters plays.

This is probably why he can irresponsibly claim that Mauer and Berkman are equivalent players.

Chairman said...

Well, the obvious initial thought is that neither Weiters nor Mauer have had an AB yet, so it's irrelevant :-)

But I don't think that it really matters to my discussion. What I'm talking about is evaluating the quality of a player, relative to what's available in free agency. That's a general question. What you may be getting at is the relative value of Mauer for a given player, which I think is slightly different.

For Rupert, once Weiters starts playing, if Weiters is awesome, and Mauer is expendable, then the additional value of Mauer for Rupert actually decreases dramatically. Think about it. Where else can he play Mauer? UTIL. At UTIL, you're comparing Mauer to players at every other position. So, his value to Rupert is no longer relative to the 13th best C, but rather the best player in free agency, regardless of position.

However, that doesn't change the value of Mauer for someone else, say Westy. Mauer's value to Westy is still the same.

So, Rup still knows that Mauer's value to Westy is the same. However, Westy knows that Mauer's value to Rup has dropped (particularly if Rup's team is deep enough such that Mauer wouldn't be one of the top 2 UTIL players). Does that mean that Westy can negotiate a better deal? If both parties are acting rationally, you'd assume yes. However, if Westy bases his willingness to pay on only his perceived value (and not on what the value of Mauer is to Rup), then all bets are off. Similar result, if Westy's affinity for Mauer makes him act irrationally.