Quick break from my behind-the-scenes but hopefully soon-to-be-posted draft model work – to address one little thing.
I’ve had a couple people mention to me that my “numbers” wouldn’t have predicted the NBA finals. Well, that is somewhat true – NO metric would have come anywhere close to predicting an average game margin for the Spurs of +14.0 over the Heat. It was crazy.
However, my metric liked the Spurs before the finals, and liked them even more when looking back and “seeing” the minutes distributions for each team.
I’m not going to do anything crazy or fancy here, I’ll keep it simple – based SOLELY off the regular season, so that no one accuses me of results bias using the finals stats as part of my data. The regular season HoopsNerd rating for the Spurs as a team was 107.9. The Heat was at 104.9. If one were given ONLY the final overall point totals for the series (986 total points scored), the prediction would have been Spurs 500 points over 5 games to the Heat 486 points (in essence, Spurs +2.8 per game). The formula to figure this is very simple, Spurs projected points = 107.9/(107.9+104.9)*986. Heat, 104.9/(107.9+104.9)*986.
Now, I suspected (but never posted here, NBA draft focused still) that the Spurs had a bigger “upside” than the Heat going into the finals, considering oft injured (during the regular season) but very important contributor Manu Ginobili would play many more minutes in the finals, thus improving the Spurs team rating. Well, here’s how it worked out for each team:
Again, I am using regular season HnI (HnI, as I always mention, is the best “predictive” rating, ignoring games missed). One can easily check my work by looking at my 2014 regular season player ratings here. Anyway, when on the court, my ratings from the regular season (HnI above) had LeBron as easily the star going into this series, followed by Duncan, then Wade & Kawhi at almost a dead heat for 3rd best, then Ginobili-Parker-Mills before you get to Bosh. The Heat had the STAR, but the Spurs had such great quality DEPTH.
Those not really in the know seem surpised about Kawhi Leonard stepping up and becoming finals MVP – but he was easily a top 30 player this last season – 23rd in WS/48, & 27th in both HnI & HnR. Only 20 other players were better in all three of those metrics than Kawhi. In fact, he and Bradley Beal were the next two that fell just outside my current 15 NBA’ers for the future. His and EVERYBODY ELSES projected future NBA WAR ranking and such can all be found here.
I digress. Predicting the finals. Back to the chart above, all one has to do is multiply the players’ HnI rating with his minutes played, compile it all, and then divide by team minutes (in this case 1200) to get a team rating. The Spurs come out to a 112.5, while the Heat are at 108.4. So, looking at player playing time allotments, The Spurs have a projected 4.2% increase from their already great regular season rating (112.5/107.9), while the Heat have a 3.4% increase (108.4/104.9).
Using the new team ratings (and total points for the series), we now come out to a 502 points for the Spurs, & 484 points for the Heat. Not much change, but a 4 point overall swing in 5 games brings the predicted per game outcome from +2.8 for the Spurs to +3.6 – which has some significance for bettors out there.
So, now, when people ask what my ratings predicted for this finals – I have a post to link to right here. Next season – with a fully formed draft model in tact (not having to create one next season) – I SHOULD be able to do some actual playoff series and projections and such, hopefully something even more in depth than I did for almost every game of the NCAA tourney.