Projecting Pelicans Wins

[greewe]

Note: This post is a crude way of projecting our wins for next season. Please don't take it too seriously!

Inspired by Nate Silver's win projections if teams added LeBron James, I decided to take a stab at projecting the Pelicans' wins using our new (nearly complete) roster.

I used a similar basic methodology as Silver, using WS/48 instead of SPM. Basically, 2014-2015 WS/48 were projected using SPS. It takes into account the last 3 years of data, and applies that to an aging curve. In the case of our rotation players without 3 years of experience (AD and Rivers), I used last year's WS/48 as a baseline before applying the aging curve, as I just can't see AD or Rivers significantly regressing from last season.

Finally, I used MM's minute projections. I assumed no injuries, because significant injuries make any sort of projection useless.



This particular model projects a 42-40 record this coming season. Obviously this is just a VERY crude method, and there are a thousand scenarios in which we beat this projection (another year of significant improvement from AD, for example), and a thousand in which we fail to meet this projection (injuries, for example).

Any thoughts? Does this seem accurate? Too low? Too high?

[MichaelMcNamara]

Yeah, I personally think AD's wins get closer to 18-22. Also, Gordon is a huge unpredictable variable. His WS/48 over the last 3 years are basically nonexistent, so if he has a contract year, that number for him could easily double. And throw a win or two extra in there for Rivers, as his rookie year greatly drags down his projection.

Side Note: Greewe, you are easily one of my favorite posters. Almost never anything negative about other posters. Just great, thought provoking facts and conversation. Thanks.

[Kumar]

Interesting projection, however, I am hesitant to use WS/48. WS are a function of how many wins your team actually gets. On a team with less wins, there are less WS to go around. And because it is derived from the box score, it is somewhat limited in projecting actual wins. We won 34 games last year with a plethora of injuries. What would have been the stats if we had won 40 and everyone was healthy?

I used a simpler way to project wins. It is based of the differential between ORTG and DRTG. They have been found to have a .934 correlation with wins. Using the formula Wins= 2.5403*Diff+ 40.908, you can get a rough idea of how many wins to expect. For example, last year our ORTG was 107.2 and our DRTG was 110.1. Using the formula we were expected to have 33.54 wins, rounded up to 34. So it was right on the money with our case.

Now lets consider Asik's defensive impact. To make things easier, I just looked at his career on/off stats. Over his career, when Asik is on the floor, he improves his team's defense by a net 4.3 points per 100 possessions. So by very conservative estimates, assuming our offense stays the same but our defense improves 4.3 points simply the virtue of having Asik we are expected to have the same 107.2 ORTG but a 105.8 DRTG. Plugging these numbers into the formula above gives you 44.46 wins.

Obviously, this is simplistic at best and doesn't account for our team improving simply due to health and chemistry, but it is a promising outlook.

[Tomdda]

All of this is fun to read, it really is. A good diversion. But I wish the season were starting tonight, sigh.

All in good time I reckon.

[WhoDatPelican]

Nice post!
So based on these projections we'd more than likely miss the playoffs again?
Round 1... I miss thee, playoffs "come home" to NOLA. (please?)

[GeauxSaints53]

So basically this:

[WhoDatPelican]

That gif is fantastic.

[Kumar]

Nice post!
So based on these projections we'd more than likely miss the playoffs again?
Round 1... I miss thee, playoffs "come home" to NOLA. (please?)

They are crude projections at best. Lots of variability that can't be accounted for, both on our team and our opponents's.

[GeauxSaints53]

That gif is fantastic.

Thanks! I went in full Pelican gif mode last night. I have another one that's more FA-related I should have done tonight.

[lsutigers33]

that is funny. lol

I can't see how we don't win 48 games. Let's hope we are healthy.

[All Metro]

I got us between 44-48

[aas3452]

The west is a beast and the east is slightly better. 45 wins may be enough to get us a spot in the playoff.

[ItIsRynosTime]

So basically this:

Yeah saw that earlier, that's funny! Very nice work, you should make more. That's more fitting for last year though.

[ItIsRynosTime]

If you want to make a serious projection, try to find a way to apply Real Plus Minus to W/L. It could take a while.

I'll give you my highly advanced projection based on the information that is in my head, which is going to be more accurate then someone using the wrong advanced statistics to try to project things that can't be projected with those statistics.

So here we go ... the Pelicans go 51-31. The formula I used for this is called educated guess. I do have some background in sports betting and handicapping.

[Kumar]

lol

[Noggy]

If you want to make a serious projection, try to find a way to apply Real Plus Minus to W/L. It could take a while.

I'll give you my highly advanced projection based on the information that is in my head, which is going to be more accurate then someone using the wrong advanced statistics to try to project things that can't be projected with those statistics.

So here we go ... the Pelicans go 51-31. The formula I used for this is called educated guess. I do have some background in sports betting and handicapping.

Wait, 51-31 and we need to fire the coach?

[ItIsRynosTime]

Wait, 51-31 and we need to fire the coach?

Yeah because we should have gone 54-28 with Mike D'Antoni in his first year with the team; those 3 less wins could cost us favorable seeding and home court in the 1st round.

[greewe]

Yeah, I personally think AD's wins get closer to 18-22. Also, Gordon is a huge unpredictable variable. His WS/48 over the last 3 years are basically nonexistent, so if he has a contract year, that number for him could easily double. And throw a win or two extra in there for Rivers, as his rookie year greatly drags down his projection.

Side Note: Greewe, you are easily one of my favorite posters. Almost never anything negative about other posters. Just great, thought provoking facts and conversation. Thanks.

I appreciate it. I always enjoy reading your stuff as well. I think AD's potential for improvement alone could catapult us from average to playoffs.

Interesting projection, however, I am hesitant to use WS/48. WS are a function of how many wins your team actually gets. On a team with less wins, there are less WS to go around. And because it is derived from the box score, it is somewhat limited in projecting actual wins. We won 34 games last year with a plethora of injuries. What would have been the stats if we had won 40 and everyone was healthy?

Thanks for the feedback. I get what you're saying, but PM variations (SPM, RPM, etc.), DRtg, and most of the "all-encompassing" advanced stats end up reflecting quite a bit of team performance as well. One of the hopes of using 3 years' worth of data is to potentially reduce some of that error. Most of the guys on that list have data from old teams (Asik even goes all the way back to Chicago!).

Over his career, when Asik is on the floor, he improves his team's defense by a net 4.3 points per 100 possessions. So by very conservative estimates, assuming our offense stays the same but our defense improves 4.3 points simply the virtue of having Asik we are expected to have the same 107.2 ORTG but a 105.8 DRTG. Plugging these numbers into the formula above gives you 44.46 wins.

Wouldn't Asik have to play all 48 minutes to get that whole 4.3 points per 100 possessions adjustment? If he plays 28 out of the total 48 minutes per game (58%) and his backup plays at last year's level, let's do a VERY simple adjustment and say the actual expected impact will be about 2.5 points per 100 possessions. Using the formula from above, that puts us at almost exactly 40 wins.

If you want to make a serious projection, try to find a way to apply Real Plus Minus to W/L. It could take a while.

I'll give you my highly advanced projection based on the information that is in my head, which is going to be more accurate then someone using the wrong advanced statistics to try to project things that can't be projected with those statistics.

Nate Silver acknowledges in the article from the OP that he could've just as easily used win shares:

"SPM has a good track record when it comes to predicting how teams will be affected by roster changes. But I’m hoping not to get sidetracked into a debate about which is the best “all-in-one” NBA metric.4 In this case my choice is pragmatic: Whereas some metrics like win shares value the very best NBA players as worth 15 to 20 wins per season, others like PER estimate that they’re worth 25 to 30 wins per season. SPM, which can also be translated into win totals, comes down somewhere in the middle, and puts James in the low 20s."

Nice try though.

[ItIsRynosTime]

Okay greewe, so you're saying 42-40, and I'm going a very positive 51-31. We'll see who is closer. I could have said 49-33 and I think I would be closer than your mathematical prediction > 80% of the time. I'm being positive here that the players are going to be at full strength and ready to go. So I'll put my "things have to go right" prediction of 51-31 prediction up against your 42-40 projection. Btw, if you're interested in making a friendly $100 wager in September of who projects closer to the win total with these same numbers. Let's see how healthy the squad is in September, I would probably make the bet either way but wait until September to make it. You would have this mathematical system backing your bet. I just threw out a number off the top of my head.

[greewe]

Okay greewe, so you're saying 42-40, and I'm going a very positive 51-31. We'll see who is closer. I could have said 49-33 and I would be closer than your "mathematical" prediction > 80% of the time. I'm being positive here that the players are going to be at full strength and ready to go. So I'll put my "things have to go right" prediction of 51-31 prediction up against your 42-40 projection. Btw, if you're interested in making a friendly $100 wager in September of who projects closer to the win total with these same numbers. Let's see how healthy the squad is in September, I would probably make the bet either way but wait until September to make it. You would have this mathematical system backing your bet. I just threw out a number off the top of my head.

Calm down man. Like I said, don't take this too seriously. I'm optimistic that we'll have more than 42 wins as well.

[ItIsRynosTime]

Calm down man. Like I said, don't take this too seriously. I'm optimistic that we'll have more than 42 wins as well.

Just wanted to see if you were confident in it enough to lay any money on it. When I am confident or not so confident in something I put money on it.

Interestingly, your win shares has Ryno #2 on the team. So bringing your second most important win share player off the bench for say 6 minutes less per game. That is 82 x 6 = 492 less minutes over the course of a season. Seems like the wrong strategy, any way you look at it, but that is what people are expecting and what posters on this forum are pining for.

[RaisingTheBar]

Okay greewe, so you're saying 42-40, and I'm going a very positive 51-31. We'll see who is closer. I could have said 49-33 and I would be closer than your "mathematical" prediction > 80% of the time. I'm being positive here that the players are going to be at full strength and ready to go. So I'll put my "things have to go right" prediction of 51-31 prediction up against your 42-40 projection. Btw, if you're interested in making a friendly $100 wager in September of who projects closer to the win total with these same numbers. Let's see how healthy the squad is in September, I would probably make the bet either way but wait until September to make it. You would have this mathematical system backing your bet. I just threw out a number off the top of my head.

First line of the original post "Note: This post is a crude way of projecting our wins for next season. Please don't take it too seriously!"

Jeez man.

[JJackisangry]

It all depends on health. If this team doesn't have significant injury problems I will go 50. I am also realistic and realize that things can go incredibly wrong, so I don't expect this necessarily. There are just too many factors to consider with such a young team and a young coach.

[HalfBreed]

A healthy Pelicans Squad will be 54-28

[greewe]

Just wanted to see if you were confident in it enough to lay any money on it. When I am confident or not so confident in something I put money on it.

Interestingly, your win shares has Ryno #2 on the team. So bringing your second most important win share player off the bench for say 6 minutes less per game. That is 82 x 6 = 492 less minutes over the course of a season. Seems like the wrong strategy, any way you look at it, but that is what people are expecting and what posters on this forum are pining for.

I'm absolutely not confident enough in this to put money on it. It was more just an exercise because I was curious what this particular method would say about adding Asik and Salmons subtracting injuries.

And in fairness to the minute projections, MM's projecting Ryno at 30 MPG and Asik at 28. Even though Ryno is coming off the bench, he'll still likely see the second most minutes of the bigs.