Years ago I started analyzing data to help make sense of past player performances, in the hope that it might offer clues about future performances. There are no guarantees, obviously. It's all about probabilities and increasingly the likelihood that what we think will happen actually *does* happen.
The more semi-definitive conclusions we can draw, the better our predictions. For example, years ago I pulled together several years of data that showed a majority of top 10 and top 11-20 players are more likely to score more fantasy points at home than on the road. So when I'm answering sit/start questions, that's one of many, many factors I draw from. Collectively, these factors present a better-than-50% probability that xyz will happen (for example, that Calvin Ridley will score more points than A.J. Brown).
There are no 100% probabilities--or even 95% probabilities, as far as I can tell. There's too much we don't know or can't know. If Myles Gaskin returns, will he return to a bellcow role, or will he be eased back with Salvon Ahmed and Matt Breida. Honestly, how can anyone know, unless Ahmed or Breida are declared inactive? But among the things we can control, we have enough tools--enough statistical history--to pull from in order to make calculated guesses on what's to come.
There are also stats that aren't easy to interpret (or maybe they are easy, but I'm not seeing it the right way). One example is the occasional randomness of DST points. We can usually get a good sense of a defense's sack or turnover potential, whether because of their own playmakers, or because of the QB they're facing. Starting an elite D against Aaron Rodgers has been risky for years. In his last three seasons, Rodgers has thrown 91 TD passes and only 10 interceptions, and this year in particular he's been largely untouched (only 18 sacks). So starting the bottom-5 Texans DST vs. Ryan Finley or even Brandon Allen makes more sense (to me) than starting the middling Titans DST vs. Rodgers. Houston might give up as many or more points, but their sack, turnover, and DST TD potential is higher.
But then there's the seeming randomness. When we look at DSTs over the last 10 seasons (not including this one), we can see that most of the top DSTs benefit more from DST touchdowns than most other DSTs. Last year three top-10 DSTs (the #4 Ravens, #9 Bucs, and #10 Jets) earned more than 20% of their fantasy points through TDs. Had they accrued a little more than the NFL average (rounded up to three DST TDs), the Ravens would have been the 8th best fantasy DST, the Bucs 12th, and the Jets 13th.
This year the Colts are the #1 fantasy DST, yet 30% of their points have come from TDs. Had they racked up a more "normal" total, they'd be no better than a high-end streamer.
On the one hand, that's not fair. Indy's deserved those touchdowns. That's part of what makes them an elite unit.
But in fantasy, I don't care about what's deserved. I care about what's predictable. And there's nothing inherently predictable about the Bucs collecting six DST TDs last year and only one this year. Removing all DST scores, they averaged 6.3 fantasy points per game last season and 6.7 this season. Last year DST TDs propped them up; this year, a lack of TDs means they scored more than seven points only four times, rendering them mostly unstartable most of the year.
That, in part, is why finding an elite DST on draft day can be tricky, and why sometimes elite-level production hinges on something that is nearly impossible to predict.
