Sleet
"busy" during Raider losses
- Joined
- Mar 13, 2009
- Messages
- 125,405
- Reaction score
- 93,989
SIS is an acronym for Sports Information Solution, which publishes a production metric (score) to compare every NFL players’ seasons. Like the AV rating published by PFR (Pro-Football Reference), SIS’s algorithms plug raw statistical data into equations and generate Total Points Saved (for defenders) and Total Points Earned (for offensive players) attributed for each player to account for actual points scored (or not scored compared to NFL averages) by teams.
The idea is to generate a value for every’s player’s season so that you can, not only compare players of the same position, but also compare players between positions. Obviously, QBs’ SIS scores are the highest. OL and NT are generally the lowest. I prefer SIS’s valuation metric over PFR’s AV rating for several reasons, one of which is it generates a higher number than AV ratings for players and appears to incorporate more advanced metrics.
Because both an AV rating and SIS production score are affected by snaps (the more snaps, the higher the number), I like to account for snaps separately. Why? Because I prefer rate statistics over raw numbers. Bosa will sack the QB more and will have a higher AV rating and SIS score when he plays 900 snaps compared to when he plays 400 snaps. (That said, a specialist’s score might not go up if he’s poor down-to-down and is forced to play full time.)
What I do is divide a player’s SIS score by his number of snaps, and report both his number of snaps and his SIS%—which is the number of points per play SIS attributes (positive or negative) to that player. I then compare that metric to PFF’s performance grade for a player, which is also evaluated play-by-play, where the number of snaps is immaterial.
You generally want players with higher SIS% and PFF grades. It breaks down roughly like this in terms of quality per play for position players (although OL and NT typically have lower SIS%):
- All Pro level > 5.0% SIS & 80.0 PFF
- Good season > 4.0% SIS & 70.0 PFF
- Average > 3.0% SIS & 60.0 PFF
- Below average > 2.0% SIS & 50.0 PFF
- Poor > 1.0% SIS & 40.0 PFF
- Bad at football < 1.0% SIS & 40.0 PFF
But sometimes SIS likes a player more than PFF. That seems to occur most often for players on good teams. Even if they play relatively poorly, they produce on the field within the framework and structure of a good team, which compensates for their limitations. The opposite is true as well. A player on a bad team can have a poor SIS% because the team sucks, which makes it harder for anyone to produce on the field, but PFF nevertheless grades the player highly.
SIS% and PFF grades provide a rough baseline to evaluate how a player has played in the past. It’s useful b/c it is hard enough to watch every Raiders’ game. No way am I watching even 6 games a week, let alone 32. And even when I watch a game, I don’t watch every individual player on each play. Thus I use these normalized metrics to provide information I don’t see myself.
Simpletons and snobs like @007 prefer to just say this player’s great or that player sucks based on what they’ve read as if they know more than GM who evaluate players every week. Or rely on raw data without context. I prefer context, enjoy evaluating numbers and use SIS% and PFF grades b/c I don’t get player evaluations each week like a GM.
Last edited:
