One of the more popular topics around the All-Star break is identifying second half studs and duds. If you are reading this hoping for another list, please accept my apology. The truth of the matter is the concept of second half players is vastly overrated. Do they exist? There is no disputing that if you analyze the statistics of some players, it appears as though there is a difference from the first half to the second half. But the problem is sheer probability dictates this will occur and it is impossible to discern a real case, it is indeed exists, from random happenstance.
A simplified manner to help debunk the myth of the second half player is to use a coin-flipping model. If 32 people flip a coin five times, one is going to flip 5 heads and another is going to flip 5 tails. While the analogy is not perfect, an argument can be made that about 1 in 32 players will have a better first half five years in a row and another of the 32 will have a better second half five straight years. The odds of flipping 4 heads and 1 tail (or the reverse) are 5 of 32. It is a fairly safe assumption that if a player has a better second half 4 out of 5 years, he will conventionally be considered a second half performer. So that means due to sheer probability, 6 of 32, or a bit below 20 percent of the player population will appear to be a second half stud and another 20 percent will have the perception of being a second half dud. With this in mind, let us go back to the guy that flipped 5 heads in a row. The chance his next flip is heads is 50/50. This, of course is the same chance a player that has had 5 straight better second halves has in doing it a sixth time.
Some people are willing to label someone as a second-half player if they have done it for only two consecutive years. Using the same model as above, 1 in 4 players will have a better first or second half.
As suggested, there are players that had enjoyed superior second halves for multiple years. And there actually may be a tangible reason for this to have occurred. But unless there are more than 25 percent of the player pool having better second halves for 2 straight campaigns, each instance can be chalked up to random distribution.
Something else to keep in mind is often the surface statistics of a player appear better or worse each half, but the underlying skills were in fact consistent throughout the entire season. There is a certain degree of luck associated with every statistic. A screaming liner can be caught or a dribbler can have eyes and sneak through the infield. By this time, most every fantasy enthusiast has figured out that one lucky hit every other week adds 20-30 points to a batting average of a full time player. It is commonplace to understand runs and RBI are team related statistics, not completely in a player's control. Even stolen bases involve a player reaching base at an opportune time to steal. With respect to pitching, wins are the proverbial crapshoot while ERA and WHIP have associated elements of luck. The primary point is the common categories used in fantasy scoring often are not a perfect reflection of a player's underlying skills.
Now let's take a look at the splits of a couple of players that are considered second half targets. Remember, even if the skills are demonstrably different, there is no guarantee they will follow suit this season.
Recently, A.J. Burnett has earned the reputation of improving in the second half. In 2007, his pre-break ERA was 4.31 as compared to 3.01 post-break. Last season it was 4.96/2.86 so for 2 consecutive seasons, his ERA is markedly improved after the break. In 2007, his pre/post break K per 9 was 10.2/8.9 so his K-rate was better the first half. His BB per 9 splits were 3.8/3.3 so he walked slightly fewer the second half, so far the skills are a wash. The big difference was allowing appreciably more homers the first half as Burnett's HR per 9 splits were 1.4/1.0. The problem is that of these 3 skills, HR per 9 involves the most degree of luck, so this difference cannot be completely attributed to better pitching, especially considering how close the other peripherals were over each half. However, last season, Burnett indeed displayed better skills after the Fall Classic. His K per 9 went from 8.9 to 10.0, his BB per 9 improved from 4.0 to 2.8 while his HR per 9 stayed almost the same, .78 to .77. Better control and missing more bats resulted in a drop in ERA of over 2 runs. Based on this, many will target Burnett for the second half. Acquiring Burnett is not the mistake. Overpaying just because he had a better second-half split last season is not the best reason.
At the end of the 2006 season, Jimmy Rollins went on a tear, carrying a consecutive game hitting streak over to the following season. His post-break decimals were .298-.346-.540, which bested .259-.323-.421 to begin the season. But his underlying skills of contact rate and plate discipline were actually better the first half. Before the break, he fanned once every 9.1 as compared to 8.1 after. To begin the season he walked once every 11 at bats and once every 13 to end the season. In 2007, Rollins had a better second half which was supported by improved skills after the break. But last season, his splits were virtually identical with first half decimals of .274-.340-.438 and second half .280-.359-.436. After two straight better second halves, he was the same player all season, meaning banking on a second half surge the season because he did it in 2006 and 2007 is not wise.
To sum up, if either Burnett or Rollins excel the second half, that does not disprove the contention that second half studs or duds is more myth than reality. Their inclusion was primarily to illustrate the connection between skills and resulting performance. The take-home message is focus on a player's skills and do not put any credence in half season splits. Even if the player does play better after the break, it is not possible to discern those from those that do so due to random probability.