The Mike Trout MVP Discussion

Foregone conclusions as recently as two weeks ago have been upended. The American League wild  card race will go down to the last day absent dramatic performances (read my thoughts here) and the AL MVP race has become very compelling for two reasons:

1. Miguel Cabrera has cooled off significantly in the past two weeks, batting .136 with 0 home runs in September (all data through Wednesday, September 11th)

2. Mike Trout, already having an outstanding year is having an explosive September, batting .417 with an OPS of 1.118. He achieved that OPS without hitting a single home run, but with an otherworldly .563 OBP.

I'm already on record stating I believe Miguel Cabrera will win the MVP, but I didn't know Cabrera was going to get dinged up and might be slow in recovering any more than I knew Trout was going to up the ante on the season he was already having. I'll also NOT bury the lead--I STILL think Cabrera will win the MVP, but that doesn't mean an analysis between Cabrera and Trout isn't worth undertaking.

My  primary measure is adaptations of Baseball Reference (B-R) and FanGraphs (FG) WAR values for both players:

I'll do a more in-depth discussion of the differences between the B-R and FG values in the offseason, but will define what I THINK these values mean. I'll state up front I have no idea how they're calculated but I don't have to know to interpret them.

Rbat is the number of runs the formulae state Cabrera and Trout delivered with their hitting--the differences in the values is far less important than the fact both sets of equations have Trout and Cabrera delivering similar runs. Trout hits for far less power than Cabrera (as does anyone else not named Chris Davis or Paul Goldschmidt) but makes up for it by leading the majors in hits. B-R Runs Created shows Trout creating 144 runs and Cabrera a major league-leading 146--essentially a wash. 

I like a simple little metric that shows the runs a player created:

Runs  +  RBI  - Home Runs

For Trout and Cabrera it looks like this:

Over 70% of Trout's plate appearances were in the #2 spot in the lineup, not typically a RBI slot. I use this number to see if top-of-the-lineup players overcome the lack of RBI opportunities by scoring runs themselves, and in Trout's case he has, but Cabrera has had a remarkable season--driving in runs AND scoring them as well. It's an idea I'll return to later.

Tne next two WAR factors, Rbaser and Rdp measure how well the players perform on the base paths. Rbaser measures stolen bases and caught stealing, advancing on passed balls or wild pitches, taking an extra base (i.e., going from 1st to 3rd on a single) and reaching on errors. All these items are measures of not only speed but savvy and it should surprise no one Trout receives value in this category and Cabrera doesn't. Rdp is how well players stay out of a double plays, another proxy measure of speed, and these two measures taken together give a thumbnail as to how well a hitter does once he actually gets on base.

The next two factors measure defensive capabilities. The first, Rfield is what it seems, how well the player actually fields his position. I hesitate when I discuss any player's fielding because fielding metrics (with apologies to John Dewan) are downright archaic compared  to the advances made in measuring offense. Trout's fielding percent isn't bad but his UZR/150 is down significantly from 2012.

Rpos is an attempt to normalize WAR so that cross-position comparisons can be made, very much like I'm doing here. Cabrera is a third baseman, a position the modern era considers an offensive position in which good defense is a nice bonus but not crucial. Center field, on the other hand, is considered a critical defensive position and rightly so since balls that aren't flagged down become extra-base hits. Therefore, the Rpos number is simply that, a number assigned to a player based on his position (and the percentage of innings he plays at that position). This is how both WAR calculations value each position:

To use an extreme example, compare David Ortiz to Joe Mauer, the catcher FG has with the highest WAR (5.0)--right from the start Mauer would have an approximate 2.5 points of WAR advantage over Ortiz just because of the positions they play and NOT due to any other hitting, base running or defensive factor. I understand why this is done, but are MVP voters really going to say "Well, that Trout didn't hit as many homers as Cabrera?" I doubt it, and if they're smart enough to be entrusted with an MVP ballot they should be able to account for differences in position when making judgments. They seemed to be able to do that in 2008 when Dustin Pedroia won with offensive numbers that were good but not stunning--but REALLY good for a second baseman.

The last value, Rrep is the familiar value over a replacement player. I'll include both B-R and FG exact definitions:

B-R--Rrep is the value of an average player over a replacement player given the player's playing time. Replacement level is set at around a .320 team W-L percentage. AL's is 22 runs per 650 PA and NL's is 18 runs per 650 PA. A player's PA is the smaller of actual PA and 4PA/G*G in order to not overvalue leadoff hitters.

FG--Replacement Runs set at 20 runs per 600 plate apperances

Allow me to distill these definitions down--these players are BAD BAD BAD BAD BAD. Be sure you haven't eaten before you see this list--it shows recent players with 600+ PA and a WAR less than 0:

Player	Year	PA	WAR/pos	Age	Tm	G	AB	R	H	2B	3B	HR	RBI	BB	SO	GDP	SB	CS	BA	OBP	SLG	OPS
Alejandro De Aza	2013	619	-0.7	29	CHW	139	558	77	147	26	4	15	57	44	133	8	19	7	.263	.320	.405	.725
Starlin Castro	2013	633	-0.5	23	CHC	144	598	55	144	30	2	9	39	26	116	20	8	6	.241	.280	.343	.623
Michael Young	2012	651	-2.0	35	TEX	156	611	79	169	27	3	8	67	33	70	26	2	2	.277	.312	.370	.682
Jeff Francoeur	2012	603	-2.3	28	KCR	148	561	58	132	26	3	16	49	34	119	14	4	7	.235	.287	.378	.665
Rickie Weeks	2012	677	-0.4	29	MIL	157	588	85	135	29	4	21	63	74	169	9	16	3	.230	.328	.400	.728
Carlos Lee	2012	615	-0.2	36	TOT	147	550	53	145	27	1	9	77	58	49	13	3	0	.264	.332	.365	.697
Delmon Young	2012	608	-0.8	26	DET	151	574	54	153	27	1	18	74	20	112	20	0	2	.267	.296	.411	.707
Jason Bartlett	2011	618	-0.1	31	SDP	139	554	61	136	22	3	2	40	48	98	13	23	10	.245	.308	.307	.615
Danny Valencia	2011	608	-0.3	26	MIN	154	564	63	139	28	2	15	72	40	102	15	2	6	.246	.294	.383	.677
Casey McGehee	2011	600	-0.9	28	MIL	155	546	46	122	24	2	13	67	45	104	19	0	3	.223	.280	.346	.626
Juan Pierre	2011	711	-0.7	33	CHW	158	639	80	178	17	4	2	50	43	41	7	27	17	.279	.329	.327	.657

Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 9/12/2013.

Add all these factors up and both Trout and Cabrera end up adding around 80-90 runs with their overall performance. Since sabermetricians state that about 10 runs equals one win, that's how both Trout and Cabrera arrive at their WAR values. FG adds another fun twist by assigning a dollar value to the performance of a player, which is the last column--any time a player delivers $35+ million in value, they're having a very good year.

As long as there are alternate versions of WAR formulae used there will be arguments over which is more accurate. I don't care--I use them not to see that Trout has a 9.4 or 9.9 WAR depending on which site you prefer, but to understand two things:
1. Either value is VERY special--both sites consider WAR values much over 7 to be historic
2. It's special because NO ONE ELSE DOES IT. I don't care what the number is, I care what it tells me, that it's unique and uncommon.
But there is a major difference, one that accounts for most of the discrepancy between the two sites. FG has Trout down as a positive fielder, B-R as a negative. As a result, the difference between Trout and Cabrera in fielding is magnified--both have Cabrera as a terrible fielder (because he is), but B-R has that account for about .6 WAR points, where FG has it at around 2.2 points of WAR, both around 2/3rds of the WAR difference between the two.

Defense is important, but is it THAT important? In both player's case (and just about every other player this side of Ozzie Smith or Mark Belanger) their offense accounts for around 75% of their WAR value, and this is completely logical for reasons I've described before and will again. Cabrera's offense is a primary reason for the Detroit pitching staff success--three of the top eight MLB pitchers in run support are Anibal Sanchez, Max Scherzer and Justin Verlander, and Cabrera hasn't had much help from the rest of the Tigers lineup in providing that support.

And with that I leave the realm of objective analysis and wade into the murky waters of opinion, a place I have no problem visiting but very rarely share due to my belief that no one really cares what I think, including me. What I attempt to do in this blog is put out the facts to allow people to reach their own conclusions. 

Winning matters. I understand it's not Mike Trout's fault that Josh Hamilton and Albert Pujols had disappointing seasons for any number of reasons or that the Angels foolishly traded Jean Segura as part of the 2012 trade for Zack Greinke or that Joe Blanton is 2-14 and other than C.J. Wilson their starting pitching has been average at best. It's not Trout's fault the Angels are stunningly weak at short and third. It IS due to Cabrera that the Tigers have been able to overcome so-so offensive production from the rest of the lineup and that the pitchers have been able to outperform what cold hard numbers suggest. Trout couldn't overcome his team's shortcomings--Cabrera did.

Put Trout in the National League and he wins the MVP hands-down since even though I've picked Andrew McCutchen, part of that is subjective since his numbers aren't great. Joey Votto has surged, but if Trout were a Diamondback, a team that isn't bad but won't win the playoffs, I could comfortably vote for him as MVP, just as I'm sure Paul Goldschmidt will receive significant MVP support. But I can't do it in the AL--the year Miguel Cabrera is having is nothing less than historic, and any deficiencies caused by his lack of speed or defense are far less than people think. Baseball is NOT 1/2 offense and 1/2 defense in terms of value, which WAR demonstrates dramatically--it's more like 80/20 tilted in favor of offense. For that reason, while it may not be FAIR to Trout, I understand and support any voter who puts the MVP check mark next to Miguel Cabrera's name on the ballot.

And if Mike Trout wins, you won't hear a peep from me. I'll understand it and even agree with the assessments, since they're down there in black and white with measures in which I have great faith. "What does Mike Trout have to DO?" the people will wail in unison if Cabrera wins--my answer is NOTHING. It's OUR job to educate everyone into how well he's playing and take the time to explain not only the makeup of the WAR equations but why they're relevant. Until that happens, it might be a long time before Trout stops being the bridesmaid...

The Historic Nature of Miguel Cabrera

It's becoming difficult to find superlatives for the play of Miguel Cabrera, but I discovered a couple last week while listening to the Boers and Bernstein Show on Chicago's 670 The Score last week. Afternoon hosts Dan Bernstein (@dan_bernstein) and Terry Boers were discussing the historic nature of not only this season but of his career and were curious about two things.

Terry first discussed an item in Sports Illustrated (I think) that looked at Cabrera's numbers for this year and asked how often in baseball history players had a .360 average, 40+ home runs and 120+ RBI. Anytime arbitrary cut lines are introduced lists can be odd and intriguing, but in this particular case, we can already guess we're probably going to see some pretty good players:

	BA	HR	RBI	Year	Age	Tm	Lg	G	PA	AB	R	H	2B	3B	BB	SO	SB	CS	OBP	SLG	OPS
Larry Walker	.366	49	130	1997	30	COL	NL	153	664	568	143	208	46	4	78	90	33	8	.452	.720	1.172
Babe Ruth	.376	54	137	1920	25	NYY	AL	142	616	458	158	172	36	9	150	80	14	14	.532	.847	1.379
Babe Ruth	.378	59	171	1921	26	NYY	AL	152	693	540	177	204	44	16	145	81	17	13	.512	.846	1.359
Babe Ruth	.393	41	131	1923	28	NYY	AL	152	697	522	151	205	45	13	170	93	17	21	.545	.764	1.309
Babe Ruth	.378	46	121	1924	29	NYY	AL	153	681	529	143	200	39	7	142	81	9	13	.513	.739	1.252
Babe Ruth	.372	47	153	1926	31	NYY	AL	152	652	495	139	184	30	5	144	76	11	9	.516	.737	1.253
Babe Ruth	.373	46	163	1931	36	NYY	AL	145	663	534	149	199	31	3	128	51	5	4	.495	.700	1.195
Mike Piazza	.362	40	124	1997	28	LAD	NL	152	633	556	104	201	32	1	69	77	5	1	.431	.638	1.070
Chuck Klein	.386	40	170	1930	25	PHI	NL	156	721	648	158	250	59	8	54	50	4		.436	.687	1.123
Rogers Hornsby	.401	42	152	1922	26	STL	NL	154	704	623	141	250	46	14	65	50	17	12	.459	.722	1.181
Todd Helton	.372	42	147	2000	26	COL	NL	160	697	580	138	216	59	2	103	61	5	3	.463	.698	1.162
Lou Gehrig	.373	47	175	1927	24	NYY	AL	155	717	584	149	218	52	18	109	84	10	8	.474	.765	1.240
Lou Gehrig	.379	41	174	1930	27	NYY	AL	154	703	581	143	220	42	17	101	63	12	14	.473	.721	1.194
Lou Gehrig	.363	49	165	1934	31	NYY	AL	154	690	579	128	210	40	6	109	31	9	5	.465	.706	1.172
Jimmie Foxx	.364	58	169	1932	24	PHA	AL	154	702	585	151	213	33	9	116	96	3	7	.469	.749	1.218
Norm Cash	.361	41	132	1961	26	DET	AL	159	673	535	119	193	22	8	124	85	11	5	.487	.662	1.148
Miguel Cabrera	.360	42	128	2013	30	DET	AL	122	551	470	93	169	25	1	75	80	3	0	.450	.685	1.135

Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 8/26/2013.

This is an easy search with the Baseball-Reference Play Index Feature and this is a very interesting list. 

FULL DISCLOSURE--this table was generated using data through Sunday, but his average dropped to .359 after Monday's game.

There are the old-school Lively Ball (it wasn't) Era hitters like Babe Ruth, Lou Gehrig, Rogers Hornsby and Chuck Klein, Hall of Famers all and deservedly so, but it needs to be understood just how much offense increased in the 1920s and 1930s--this chart shows the batting average leap beginning in 1920 through around 1941 when World War II affected everything from available players to materials to make baseballs:

The National League average batting average was .303 in 1930, a full 50 points above what it will likely be for both leagues in 2013. In 1930 Chuck Klein hit .386 when the league batted .303, or about 27.4% better than the league. Cabrera's .359 compared to the AL's .256 is 40.2% better--Cabrera is hitting over 100 points better than the league. Some day I'll follow up on THAT to see how often it occurs, but I suspect it won't be often.

Since this is a small list I can use advanced metrics from FanGraphs to give a fuller picture. This chart adds two very useful numbers, Weighted On-Base Average (wOBA) and Weighted Runs Created Plus (wRC+) for the 17 seasons in the chart above:

wOBA is a Tom Tango creation that moves beyond traditional hitting metrics by  measuring value they deliver instead of just counting them. Scaling it to resemble traditional hitting metrics helps to understand how good a season a player had. Suffice it to say that any season with an wOBA of .400 or higher is excellent, which makes these 17  seasons simply stellar. 

wRC+ is another Tom Tango creation that allows comparisons of offensive production across parks and eras. Any time there is a "+" after a number (usually OPS+ for offensive players, ERA+ for pitchers) it means the number is normalized. If you know and understand this, skip to the next paragraph. All normalization means is how much better the player was compared to the league. For example, Cabrera has an OPS of 1.137 and an OPS+ of 202. You can see the calculation for yourself here, but in essence this states that Cabrera's OPS+ is twice as good as the league's. Here are the pertinent numbers:

If you actually follow the formula and do the math, well, I got an OPS+ of 210 vs. his actual 202, but I chalk that up to me. The League values do NOT include Cabrera's production, so that's not it--let's just leave it aside for now since the important point remains the same--his production is DOUBLE the league's average, and this is historic.


wRC+, with a "+" behind it, is another normalized stat with a league average of 100. Cabrera's wRC+ is 206, again showing that his production is twice the league average. This chart shows every season since 1901 in which a player had a wRC+ of 200 or higher:

 































These 32 seasons can be grouped:
1. Enhanced offense years from 1994-2004 (Frank Thomas excepted)
2. The absolute best hitters in baseball between 1940-1960 in Ted Williams, Mickey Mantle and Stan Musial
3. The Lively Ball hitters already discussed
4. The Dead Ball Era in which Ty Cobb and Ruth (and no one else) were able to transcend the lack of home runs and still generate tremendous offensive production. Not that I care all that much for RBI, but Cobb drove in 102 with only 6 home runs in 1917.

And then there's Cabrera, the first such season since 1957 (I'll discount Thomas's 1994 slightly because it was a strike-shortened year and the others between 1994-2004 have their well-known issues)--the first such season in FIFTY-SIX YEARS.

A season is a season, and by this point it's safe to state that Cabrera's 2013 isn't a fluke, After the discussion of these seasons, Dan Bernstein mentioned that Cabrera has averaged around .320 and 33 home runs in his last 10 seasons--how many stretches in baseball history matched this?

This is very solid company and helps put modern players in their proper historical context. I'll discuss this in far greater depth when we get closer to Hall of Fame balloting but there is a tendency to place greater emphasis on numbers for players we DIDN'T see play vs. those we did. It's a spin on the old "familiarity breeds contempt" notion that we discount and take for granted what we can see, but when the time for evaluation comes for Albert Pujols, Vladimir Guerrero, Frank Thomas and Manny Ramirez, this level of sustained excellence will have to be taken into consideration. There will be other factors as Mike Piazza already discovered and at some point baseball will have to make some kind of determination on how to handle players from approximately 1990-2007, but that's a different post. In an age where virtually anything is labeled "historic," it is very important to recognize the truly stunning when it is seen.

It used to be rare to hit 40 home runs, and around 1995 it became far more prevalent, but that incidence has shrunk again. How much this is due to reduced PED use vs. other factors won't be known for a long time, but it appears only Cabrera and Chris Davis will cross that threshold in 2013. In this day and age we can't automatically assume Cabrera will be able to continue his offensive production into his late 30s, but it's very important to remember that he's only 30 now--if he puts up 3 more solid-to-spectacular seasons (and not even equal to this year--we're entering an era where .300/30 HR/100 RBI will again be special), he'll have achieved a level of sustained excellence that is extremely rare in baseball history. 

There are good young hitters in baseball today--Paul Goldschmidt for power, Mike Trout for overall offensive production, Yasiel Puig for raw potential, but it might be quite some time before we see the likes of Miguel Cabrera again, someone who puts it all together in one package. I didn't mention WAR all that much in this post because it wasn't my focus, but it helps illustrate his defensive shortcomings--he easily had the worst fielding season of any of the 17 seasons in the chart. I'm sure there's a breakeven point where his defensive shortcomings would overcome his offensive output--here are his FanGraphs WAR values:

 

Cabrera's hitting has generated around 67 runs, his fielding has cost around 15 runs. So far this season he's made 12 errors, suggesting he'd have to make something around 60 errors or so in order for his fielding to be detrimental. He'll probably end the season with around 15-17 errors and adding another 15-20 runs with his hitting. In other words his hitting outweighs his fielding by around a 4:1 factor or so. Breaking down WAR into its components is very useful for precisely this reason, to see at what point defensive liabilities transcend hitting, but in Cabrera's case, it's not even close.

This last chart is so big that I'll finish writing and then put it in, but take the time to not just look at it but study it for what it shows--the difference between the player with the highest wRC+ and the second-highest by year. For example, Chris Davis's 182 wRC+ is 13.2% below Cabrera's, one of the largest gaps between #1 and #2 in recent years for players not named Barry Bonds. As you look at the names and differences on this list, it should become apparent how amazing Cabrera's season is, because he's accomplishing it in what is NOT a hitter's era like between 1920-1940 and 1995-2007. Pitching is on the rise and Cabrera STILL is having one of the best offensive seasons in baseball history. This is why I use numbers--not just to show something, but to hopefully illustrate how special a given season is, and these numbers aren't cherry-picked to paint a picture. There's a mix of traditional and sabermetric, and the only real question left in my mind is if any MVP voter will have the audacity to NOT use their first place ballot for Cabrera.