NBA Team Win predictions as of now vs the Lines

I love over/under columns.  They’re always great fun to read and they give us an opportunity to gauge the value of the insight being provided by the author. Does he truly have something with his process of evaluation or does he need to go back to the drawing board?

You may be aware that I’ve been building such a model for predicting NBA win totals (See here for the latest version and here for the Basics). Now, the model’s not quite done yet as rosters aren’t final, I need to add in some additional player projections and I need to adjust the minute allocations to reflect injuries and final depth charts but it’s at a point where I can look for teams that are being undervalued or overvalued by the public and the media. How do I do that? Simple enough I go look for the moneylines.

Now before we continue, I’m going to drop an explanation from wikipedia on what the moneylines are and mean :

Moneyline odds

Moneyline odds are favoured by American bookmakers. There are two possibilities, the figure quote can be either positive or negative.

Moneyline odds are often referred to as American odds. Moneyline refers to odds on the straight-up outcome of a game with no consideration to a point spread.

Positive figures
If the figure quoted is positive, the odds are quoting how much money will be won on a $100 wager (this is done if the odds are better than even). Fractional odds of 4/1 would be quoted as +400, while fractional odds of 1/4 cannot be quoted as a positive figure.
Negative figures
If the figure quoted is negative, then the moneyline odds are quoting how much money must be wagered to win $100 (this is done if the odds are worse than even). Fractional odds of 1/4 would be quoted as -400, while fractional odds of 4/1 cannot be quoted as a negative figure.
Even odds
Even odds are quoted as +100 or -100. Some bookmakers display the negative symbol while others do not.

The rest of the explanation is here. To put it in table form:

What this means in plain terms is if you’re quoted a +X line it means that if you bet $100 and win you’ll get 100 + X back and if you’re quoted a -X line it means that if you bet X and win you get X+100 back. Got it? Excellent, let’s get to the lead.

To get the over/unders I went to betus.com (see here), which currently has the bets listed and open. If I line up all the over/unders against the current predictions it looks like this:

Based on the current model, there are ten lines where you would have an advantage (you know if you were into that sort of thing).  I’d discount the Pacers, Rockets and Spurs for having extenuating circumstances (in order Collinson and being to close to the error margin, Yao, the whole Splitter/Blair minute allocation thing).

So we are left with seven lines with advantages. The Wizards line is laughable (go read this).  The Nets,Kings , Blazers and GSW are undervalued but I would stay away from the Nets because if they swing the Carmelo trade they would be valued right.  The Clippers and 76ers are undervalued but I would be nervous about Griffin who might have a Bird/Magic type of rookie season.

Finally we have five lines with clear advantages and no concerns as of now:

  • Under: Washington, Philadelphia
  • Over: Kings, Blazers, Warriors

We will revisit this again once the final model is set (and prior to the bets being closed).

21 Comments

  1. […] Basketbawful thinks the Raptors might be one of the worst teams in the league. Wages of Wins has the Dinos slated to win 25 games while Arturo Galletti is a bit more generous with a 26-win prediction. […]

  2. 10/18/2010
    Reply

    Interesting.

    Ever thought of making money this way? It’s not really “gambling” if the odds are heavily in your favour, is it?

  3. Chicago Tim
    10/18/2010
    Reply

    As I understand it, betting lines do reflect conventional wisdom because the idea is to end up with an equal number of bettors on either side of the line. The odds maker doesn’t measure accuracy based on outcomes, but rather based on balancing the bets. Thus if too much money is bet on one side, the odds will be adjusted accordingly, regardless of whether the odds maker thinks it makes sense.

    So the odds makers are not necessarily wrong about the Wizards, 76rs, Kings, Blazers, and Warriors. They probably are right that most people overestimate the first two teams and underestimate the last three — assuming you have not done the reverse.

    Nevertheless, if you can find a predictive model that beats conventional wisdom every time, well, money can be made!

    • 10/18/2010
      Reply

      That’s exactly the point of the post. The lines are meant to draw equal money to both sides and so are a good reflection of where the public perception on every team is. If you find that the public perception is predictably skewed then yes there is money to be made. The trick is to play safe bets with large margins in your favor. This is how professional gamblers make money.
      If I can build a decently accurate model then yes there’s money to be made there (jury’s still out on the model :-) )

      • Mike
        10/18/2010
        Reply

        I feel like the true “safe” bets are the under bets. Injuries can wreck havoc on the over predictions. The Wizards however are expected to be terrible and have very little to offer for trade for them to gain a 10+ win player. Philly will most likely trade away Iguodala (if a trade were to happen) for a 10+ win player (or less), so even with trades they will likely stay the same or get worse.

        • 10/18/2010
          Reply

          I tend to agree but I think the Blazers, Warriors and Kings are all fairly safe as listed. They all have higher ceiling than listed for various reasons (Oden for Port. , trading Monta for anything including a ham sandwich for the Warriors, and Cousins for the Kings). The Spurs at the over would worry me because of Splitter, age and injury.

  4. some dude
    10/18/2010
    Reply

    That Wizards line is ridiculous. 32.5? Are they adding up last season and this one?

    • 10/18/2010
      Reply

      It’s meant to draw even money but yeah, you get set that at 22.5 and still make a huge killing.

  5. 10/18/2010
    Reply

    With the Blazers’ injury history, I wouldn’t touch their number w/ a 10-foot pole.

  6. some dude
    10/18/2010
    Reply

    I agree with RG. The Blazers variability seems huge. And besides, I see them as like a 54 win team, so it’s not a big enough disparity. I could see Portland winning 47 due to injuries and such. I can’t see Washington winning 33 unless half the teams strike…and even then i’d think about it.

    I think the Raptors are a good under pick. I don’t see how they win 26 games unless they magically learn to play D. But I’ve said enough about that. Kings are tempting as an Over.

  7. 10/19/2010
    Reply

    I like the Blazers. The Raptors aren’t a bad under pick but only because they might tank (not because of their roster)

    • 10/19/2010
      Reply

      And I wouldn’t put it past the organization to do something like that. But they could also end up making a trade or three (remember, Calderon is on the way out, they still have the Bosh trade exception, and several expiring deals) which makes them a bit risky.

      • 10/19/2010
        Reply

        Note the the model hates the Raps more than any other team as an over under.

  8. ET
    10/19/2010
    Reply

    Were their similar predictions made last year? How did they do? If not, can you retrospectively do this for many past seasons and see how the team predictions did — and even better, compare to Vegas odds in years past (where would that data be?) and see these sort of arbitrage opportunities really exist?

    • ET
      10/19/2010
      Reply

      oops — “there”, not “their”! :-)

      • 10/19/2010
        Reply

        Wasn’t done for last season (this is the first go around). Once I finish the model I can compare to previous seasons ( I need to figure out where I can get opening day rosters).

  9. 10/24/2010
    Reply

    Four games? You honestly will stake the credibility of your system on the idea that the Wizards will win only 4 games this year?

    I’m curious what system corrections you will make if the team manages to quintuple your projection to a still-miserable 20 games.

    Outliers exist in all data sets, but I find them instructive. They raise questions. In this case an area I suggest you consider, to improve on Prof Berri’s work, is an age/minutes component. Rookies become veterans over time, veterans lose their edge eventually. The Wizards are banking on development by their youth corps. I’d agree that they’ll struggle, even fall short of the Vegas line, but I’d jump on the over with bags of cash if the line were set at 4.1 wins. That’s absurd… Silly indeed.

    • 10/24/2010
      Reply

      doclink,
      You are correct however the Wizards prediction is not four games. That’s the uncorrected and not final number. Go here for the full detail. It’s not the first time a team projects that low but the numbers get funny down there. Every team that low has managed to win 11-12 games. However for this article vs the lines? Take the ridiculuous under and laugh all the way home. Oh and I do include an age and minute projection in all the numbers.

Leave a Reply

Your email address will not be published. Required fields are marked *