## My Miami Heat prediction: 63.0 wins

In the midst of the “Summer of LeBron” I did a rough estimate of how many wins the Heat would produce using win production generalizations and a series of logical assumptions.  Using that method, I predicted the Heat would win 63.3 games.

Tonight I did a more precise estimate using each player’s actual win production statistics over the past three seasons combined with an estimate of the playing time each player will receive based largely upon each player’s past minutes, age, and physical condition.

Using this method I estimated the wins and losses each player would produce.  The sum total of the individualized estimates produced nearly the same result I obtained using the rough method I used last summer.  (the individualized method is the method I’ve been using throughout my “2010-11 NBA win estimate” posts — when I finish predictions for every team I will write a compilation “NBA Preview” that will link to every team and will be organized by conference and division.)

Below are the win production estimates for every player on the Miami Heat roster.  As always, my system does not claim to be prescient, but it has proven fairly accurate.  Of course, the final verdict is always left with the reader.

When reading the chart, the first number for each player is the projected wins he will produce.  The number after the underscore is the losses I project he will produce.  If there is a parenthesis around either number, that means the player is either so productive or so unproductive that he produces either negative losses (superstar) or negative wins (horrible).

### 8 Responses to “My Miami Heat prediction: 63.0 wins”

1. Chicago Tim Says:

Seems a little low unless the Heat fail to play Bosh at center and James at point, thereby depriving Haslem and Miller of minutes in favor of lesser players playing center and point. Was that part of your prediction?

2. tywill33 Says:

I just went by the number of minutes each player has averaged over the last 3 years, weighted toward the recent years.

James routinely plays above 3000 minutes. Bosh and Wade play a lower amount of minutes. Wade’s style is conducive to injury and Bosh has averaged only 70 games per season for the past five seasons. Haslem is up and down with his production and defense. Miller generally plays around 2000 minutes and produces very consistently productive results.

The rest of the roster is either crap or borderline crap. The one possibility for increased performance is Bosh. If he changes his style and becomes a dedicated rebounder and high percentage shooter, he could add a few games.

For those reasons I’m pretty confident that if all things remain constant the prediction will be +/- 3 wins.

3. Chicago Tim Says:

Thanks for the explanation.

If I were in charge Haslem and Miller’s minutes would go up, but that might not happen because they are not confirmed starters.

You make an interesting point about Wade and Bosh’s history of missing time due to injury. Maybe playing on a great team will reduce their risk of injury, but maybe not.

I’m not sure how well your method works when this team has undergone so many changes and has so many significant new players, but you do try to take into account the risk of injury, which many predictive methods ignore.

4. tywill33 Says:

Yes, a player’s season-to-season win production has a lower correlation when a player switches teams.

Last season, I did predictions for a handful of teams. My method was dead on for the Lakers, Blazers, Nuggets, and Spurs. But those were all teams with established rosters, so a monkey could have guessed their totals.

I was prouder of the fact that my predicted win totals for the Bucks and Pistons were far closer than experts predicted for the Bucks and Pistons. I was close on your Bulls, also. But I was way off on the Thunder, Timberwolves, and the Sixers.

This season I want to put the predictions out there for every team so I can see how it does. I would expect about 70% accuracy.

A little inaccuracy would actually be okay by me. Too much accuracy takes the fun out of following sports.

5. tywill33 Says:

Oh, after I’m done I have to reconcile my predictions to make them jibe with available wins. In other words, the grand total of wins predicted has to equal 1230 wins. I’m not quite sure how I’m going to do that. I’m hoping the totals will be close, but if they are that would only be accidental.

6. Chicago Tim Says:

I look forward to the final predictions.

7. arturogalletti Says:

Ty,
Awesome work as always. I love coming here to learn.