Next up in our look at what goes into the Fantasy Alarm daily projections is the influence of recent performance on the adjusted projection. It's a little surprising this isn't a hot button topic along the same lines of batter versus pitcher (BvP) analysis because they both have conclusions emanating from intuition that research suggests are suspect.
Cutting to the chase, research concludes that a hitting streak, either hot or cold, isn't predictive. At best, a streak should be used as a tiebreaker. The primary source of this is from our old friend, The Book: Playing the Percentages in Baseball. I have also conducted my own studies that corroborate these findings.
For that reason, I don't adjust batter's projections at all based on recent performance,
However, when constructing my lineup, I will look at a hitter's contact rate while his streak is occurring, Almost all hot streaks involve a higher than normal batting average on balls in play (BABIP) while the reverse is true for cold streaks. When a player is scuffling, his BABIP is usually low. While there is obviously luck involved with BABIP, a player can be making hard or soft contact within that streak. This data isn't readily accessible but contact rate is available and during stretches of hard contact, a player tends to strike out less.
In the spirit of using streaks as a tiebreaker, if a hot streak is devoid of strikeouts and the matchup is otherwise favorable, I may go with a hot player to differentiate two close players. I'll fade perceived hot players if they're striking out more than usual.
Similarly, I'll avoid cold players that are whiffing more than usual, regardless of price. On the other hand, if a hitter is cold but isn't fanning any more than normal, I have no issue taking advantage of the depressed cost.
The story with pitching is a little more interesting. The Book finds that a pitcher with a stretch of four solid games is more likely to continue his better than normal pitching while a pitcher struggling is likely to return to his baseline. In other words, hot streaks are real while cold streaks aren't predictive,
The manner I incorporate this into the daily projections is to include a weighted average of his past three weeks with his normal baseline, The weight is set such that for it to have a real impact, the three weeks of data needs to be well above normal, which can only be attained with four solid efforts,
On a personal level, because I'm already factoring recent performance in, I don't let recent performance further bias my selection when choosing my DFS pitcher. If however, you're more of a Zen lineup constructor then using a hot pitcher, especially if they're less than elite, you can feel safe using the hot pitcher.
TO REVIEW
Hitting streaks (hot and cold) should be used strictly in lieu of a coin flip if everything else is a wash
A hot pitching streak is predictive of another solid outing while a cold pitching streak in non-predictive.