# Then and Now: Cliff and CC

On Thursday, I offered a primer of sorts with regards to how players should and should not be evaluated.  The key involves using the player’s true talent level–derived from weighting the last three or more seasons of data–in order to project current numbers.  Without regurgitating the entire post I will sum it up by saying it is always incorrect to solely quote this season’s statistics when conducting such an evaluation.  At a certain point later in the year, like around now, the 90-100 game mark, current season numbers can truly garner the most weight, but still require the previous three years to serve as supporting evidence.
All of this leads me to my major point today, using math anyone and everyone can understand: Four starts in April is not 90-100 games into the season.  It may seem like common sense and elicit sarcastic reactions at the obvious nature of this statement, but rewind back to April and take another look at the almost uniform reactions to both Cliff Lee and CC Sabathia.  These were two pitchers heading in what appeared to be different directions entering the year–CC winning the Cy Young Award in 2007, primed for a big contract after the season, and Lee struggling to even receive a guaranteed spot in the rotation.  Despite this, they produced results of the inverse extremes through four starts:

• Cliff Lee: 4 GS, 31.2 IP, 11 H, 1 ER, 2 BB, 29 K, 0.28 ERA, 1.21 FIP
• CC Sabathia: 4 GS, 18 IP, 32 H, 27 ER, 14 BB, 14 K, 13.50 ERA, 7.25 FIP

Most Cleveland fans didn’t know how to react, mainstream writers wrote Sabathia off as struggling under the weight of his impending new contract, and I’m sure John Kruk cherrypicked videos to show that, without question, something was wrong with CCs mechanics.  After nothing more than four starts the minds of many were made up that Cliff Lee was absolutely amazing and Sabathia was in the midst of a severe post-award dropoff.  Recall Thursday’s example of a 3-9 hitter vs. a 300-900 hitter: with such small samples, a few more games of data could drastically change the outlook.
Granted, CC and Cliff were on extreme opposite ends of the four-start spectrum but what should have been asked is how their early season performanc affected their true talent level.  Did it at all?  Or were their starts so incredibly good or bad that they were capable of overcoming small sample shortcomings to drastically alter the true talent level?  For starters, Lee was projected to post a 4.38 FIP entering the year, with Sabathia at 3.25.
Sabathia’s 7.25 was much higher than his projected 3.25 whereas Lee’s 1.21 was about four times lower than what was projected for him.  Plugging their starts into Sal Baxamusa’s in-season Marcel projector tells us that, after these four starts, Lee was projected to post an FIP of 4.03 over the remainder of the season whereas Sabathia would post a 3.47 FIP in the coming weeks and months.  Lee improved by about one-third of an FIP point while Sabathia dropped by about one-fifth of one.  Something to keep in mind is that each of these were extreme small samples, with insanely good or bad numbers.   Still, the true talent levels had not changed as significantly as some thought.  Cliff Lee was not considered capable to be a 1.21 FIP pitcher with a 14.5 K/BB.  His projection understood that he would likely outdo his pre-season 4.38 FIP projection but not to the extent he did in the early going.  With 27-29 starts remaining after those initial four, the balance FIP of 4.03 would more than outweigh the early 1.21.
Sabathia, however, had been bad enough in the first four starts that his projection understood he HAS to perform better, but that his overall numbers this season might not be as great as was thought possible prior to the beginning.  A 7.25 FIP through four games would largely be drowned out by 30 more starts of 3.47 FIP, but the end result would not see Sabathia with a 3.25 FIP like his pre-season projection.  He would perform much better than his first four starts but not better enough to even things out to the 3.25.
All told, after these first four starts, Lee’s 4.03 projection deemed him a good, not great, pitcher, whereas Sabathia’s 3.47 still merited him some recognition in the “great” department.  What happened from that point until now?  I’m glad you asked!  Since those fateful fourth starts, Cliff Lee has made 14 starts at a 2.48 FIP while Sabathia has made 16 starts at a 2.32 FIP.  For the season, excluding last night’s action, Sabathia’s FIP at the half was 3.23, Lee’s was 2.31.  It appeared that, despite his four poor starts to begin the season, Sabathia had pitched so well in the coming weeks that his halfway point FIP actually outdid his projection, something few people thought possible following his second straight 9-run performance in April.  Lee, however, was still outdoing his projection by a large margin even though it gravitated towards Earth.
Plugging their halfway point numbers in, Sabathia is projected to post a 3.21 FIP over the rest of the season while Lee’s projection calls for a 3.66 FIP.  Suffice it to say, these projections mean that relatively nothing will have changed for Sabathia, as his FIP would actually be an improvement, ever so slightly, over his pre-season 3.25 FIP.  Lee, on the other hand, would alter his projection significantly.  His FIP would be somewhere in the upper two’s or lower three’s, which would vastly outdo not only what was thought of as probable prior to the season and possible during the season.
Entering next year, however, his 2008 numbers will be weighted the heaviest but not the sole factor in determining his true talent level.  Lee may very well post better numbers in this particular season, but that doesn’t make Sabathia’s any less remarkable.  Based on their track records, CC is more likely to sustain performance like this as well.  The point of all this is not to get carried away with small samples of early season performance.  It’s fine to look at what has happened and, on a granular level, analyze what went into those numbers, but don’t make Cliff Lee out to be the next Bob Gibson based on four tremendous starts that nobody could sustain, let alone a guy fighting for a rotation spot in spring training without a proven track record.  It’s not to say he couldn’t continue to pitch on a tremendous level or that he lacks talent, but that judgments and evaluations need to be based on concrete evidence, like a true talent level, and as I mentioned before, four starts is not the true talent level of a pitcher.
I believe it was Rob Neyer who once said that you have two choices as a baseball writer early in the season: have fun with the numbers or bore everyone about how the samples are so small.  I definitely agree, so long as that fun does not include going overboard and determining that a player has “changed” something based on four starts or something like twenty games for a hitter.