NBA Championship: Beating the Odds

[Griffin]

So your favorite NBA team made the playoffs. But don't celebrate just yet. The field was narrowed down from 30 to a sweet 16, but your team still has to win exactly 16 more games in order to be crowned this year's NBA Champion. And that is no easy task. So before you go off and bet your life's savings or your kid's college tuition fund on your favorite team winning the trophy, let's look at the (Vegas) odds each playoff team has of capturing the coveted title:

TEAM			 ODDS	PROBABILITY	PROBABILITY
				OF TITLE	OF NO TITLE

Boston Celtics		  7-4	30.29%		69.71%
Los Angeles Lakers	  2-1	26.51%		73.49%
San Antonio Spurs	 11-2	 9.64%		90.36%
Detroit Pistons		  6-1	 8.84%		91.16%
Phoenix Suns		  7-1	 7.57%		92.43%
Utah Jazz		 15-1	 3.53%		96.47%
New Orleans		 15-1	 3.53%		96.47%
Dallas Mavericks	 25-1	 2.12%		97.88%
Orlando Magic		 30-1	 1.77%		98.23%
Cleveland Cavaliers	 30-1	 1.77%		98.23%
Houston Rockets		 35-1	 1.51%		98.49%
Washington Wizards	 50-1	 1.06%		98.94%
Denver Nuggets		 60-1	 0.88%		99.12%
Tornto Raptors		100-1	 0.53%		99.47%
Philadelphia 76ers	200-1	 0.27%		99.73%
Atlanta Hawks		300-1	 0.18%		99.82%

You may or may not agree with these numbers. After all, these are just predictions. How the odds are dervied, I am not even sure. But there is one thing certain each and every year: even the best team in the league is much more likely to not win the title than it is to win it. This is true for all sports where a lot of teams are competing for the same trophy. And this year in the NBA it's even more true, because there are many good teams playing at a high level.

Yes, in a way, I am stating the obvious. But there is a reason. A lot of people seem not to realize, or appreciate, that to win the title in any given year, even the best team has to beat the odds. A good performance during regular season will improve your odds. The talent level, coaching, experience, readiness, effort and energy will improve those odds further. But there is only so much a team can do to prepare. The rest is luck.

 

[Cheesewater]

They're not even predictions really though, right? Vegas wants to make money. Odds are set to get people to bet. It really has no meaning as to who might win. If I understand betting odds that is.

 

[Griffin]

They're not even predictions really though, right? Vegas wants to make money. Odds are set to get people to bet. It really has no meaning as to who might win. If I understand betting odds that is.

Either way, the point is the same. I don't know if anyone actually calculates predictions other than those odds. But unless you predict that say the Celtics are more likely to win the title than all the other 15 teams combined, then they will still be more likely not to win the title than to win it.

 

[YouJustGotSUNSD]

I agree with the rankings. Vegas is going to be the most objective way of seeing who is most likely to take it all, as money > * when it comes to opinions

Mind you, these odds may change depending on betting. If they see a huge surge in an investment of one team over another, the odds may fluctuate later down the line. San Antonio is the most inaccurate imo, as they should be a slightly lower percentage than they are.

100 bucks on the suns gives you 700 back if they win it all. Not too shabby imo.

 

[Bufalay]

You must be registered for see images

 

[Irish]

The one "odd" part about Vegas odds making is that money tends to reflect the size of the fan base. Given a choice between the fan base of the Lakers and of New Orleans (which didn't really support a great team), the Laker base is much larger. It is hard to separate out the smart money from the fan driven money going toward certain teams.

 

[SunsFanVegas]

Actually, the Suns line at 7:1 is pretty good because it is high enough for the general public to think that they don't really have a chance at winning, but low enough for the books to pay out actual sharp bettors who bet the Suns. Living in Las Vegas, I see it all the time. It's not very often that the odds on favorite actually wins the title.

 

[elindholm]

But there is one thing certain each and every year: even the best team in the league is much more likely to not win the title than it is to win it.

I really don't think that's true "each and every year." In fact I don't even think it's true most years. In a typical NBA year, I would say that the best team has better than a 50% chance of winning the title, regardless of how Vegas sets the odds. I agree, however, that there is no 50%+ favorite this year.

I was looking over that list and trying to figure out the best value bet. Maybe Dallas? Also, how did you calculate your "probability of title"? A 2-1 underdog represents a break-even probability of 33% (minus house take). At 7-1 the Suns should be 12.5% (which is too high, of course).

 

[Cheesebeef]

does anyone realize that in the last 10 years, the team with the best record in the league only won the title twice? Or that the West's 1 seed has only been to the Finals once since 2000?

 

[SunsFanVegas]

This could get into a whole new heated discussion, so all I will say is that Vegas knows what is going on and writes the odds accordingly. They are not in the business to lose money.

 

[elindholm]

This could get into a whole new heated discussion, so all I will say is that Vegas knows what is going on and writes the odds accordingly. They are not in the business to lose money.

They aren't in the business of picking winners either. All they care about is balancing the wagering. They set the odds according to how they think people will bet, not on what they think the sporting outcome will be. You're trying to outsmart the other bettors, not the house.

 

[abomb]

Utah at 15:1 is pretty nice.

 

[Griffin]

I really don't think that's true "each and every year." In fact I don't even think it's true most years. In a typical NBA year, I would say that the best team has better than a 50% chance of winning the title, regardless of how Vegas sets the odds. I agree, however, that there is no 50%+ favorite this year.

But then that would imply that all the other 15 teams (or 29 if you look at the odds at the start of the season) have a combined less than 50% chance of winning the title. That's just seems very improbable considering how much can happen en route to a title. And what about the second-best team? If they get a 25% chance (which would already make the top team twice as likely to win the title than their biggest competitor) that only leaves less than 25% probability that any of the other 14 teams would win it. But don't we usually have at least 3 or 4 teams that are considered serious championship contenders?

Also, how did you calculate your "probability of title"? A 2-1 underdog represents a break-even probability of 33% (minus house take). At 7-1 the Suns should be 12.5% (which is too high, of course).

I just took the inverse of odds for each team and divided it by the sum of all the inverses. I realize that the betting odds are not the best numbers to use for analyzing true probabilities of a future event, but those are the only numbers available and they help to illustrate the point.

 

[Ronin]

You must be registered for see images

 

[Skumbag]

You must be registered for see images

NBA: Where side jobs happen.

 

[TucsonDevil]

You must be registered for see images

NBA: Where the 'fix' happens.

 

[elindholm]

That's just seems very improbable considering how much can happen en route to a title.

Not to me. If you look over the past 25 years, I think the eventual champion has been the expected one at least half of the time.

But don't we usually have at least 3 or 4 teams that are considered serious championship contenders?

Not in my opinion. The league and media try to prop up several teams as "serious" contenders, but it's usually empty hype. Typically I think there are only two or three teams that have more than a single-digit probability of winning. This year is different, however.

I just took the inverse of odds for each team and divided it by the sum of all the inverses.

The math is wrong. I think you treated x:y odds as a y/x chance, but it's not, it's y/(x+y). For example, 3:2 odds are 40%, not 67%.

 

[Griffin]

Not to me. If you look over the past 25 years, I think the eventual champion has been the expected one at least half of the time.

Not in my opinion. The league and media try to prop up several teams as "serious" contenders, but it's usually empty hype. Typically I think there are only two or three teams that have more than a single-digit probability of winning. This year is different, however.

So, any given year, if you had to bet on one of the following events happening, which one would you put your money on?

• The best team (your choice) winning the title.
• One of the other 15 teams winning the title.

Would you really pick the first option? Imo, the "best" team is never that good that they are more likely to win the title than all the other teams combined. Even if they were, too much depends on luck. It is far more likely that a single team will have bad luck (lose a player) preventing it from winning the title, than all the other teams suffering similar fate at the same time.

The math is wrong. I think you treated x:y odds as a y/x chance, but it's not, it's y/(x+y). For example, 3:2 odds are 40%, not 67%.

I did not do y/x. And your approach results in combined odds for all teams at greater than 1. What I did was take 1/(x/y) for each team and then divide that by the total so that each team has a share of total probability of 1. But either way you do it, no team will end up with better than 50% chance of winning the title.

 

[elindholm]

So, any given year, if you had to bet on one of the following events happening, which one would you put your money on?

• The best team (your choice) winning the title.
• One of the other 15 teams winning the title.

Would you really pick the first option?

In about half of the years, yes, maybe a little more often than that. Note that we are talking about placing the bet at the start of the playoffs, not the start of the regular season.

I did not do y/x. ... What I did was take 1/(x/y) for each team

1/(x/y) is the same as y/x.

And your approach results in combined odds for all teams at greater than 1.

Yes, I understand that. What I'm talking about is which numbers you use before you scale the sum to 1. To make your system work, you need to use y/(x+y) and then scale those.

For example, a team at 5:1 odds is (in the judgment of the oddsmakers, or more accurately the projected betting population) half as likely to win as a team at 2:1 odds. Here's how you can see that: consider six cases for each team. To break even, the 5:1 team will win one time out of six and lose the other five, so that's a 1/6 probability. The 2:1 team will win twice out of six and lose the other four, so that's a 1/3 probability.

By your way of calculating, the 5:1 team is only 40% as likely as the 2:1 team to win (1/5 = 1/2 * 40%). That's why the relative weights of your numbers are all wrong.

 

[Griffin]

Yes, I understand that. What I'm talking about is which numbers you use before you scale the sum to 1. To make your system work, you need to use y/(x+y) and then scale those.

Ok then, here it is:

TEAM			ODDS		PERCENT	SCALED

Boston Celtics		7	4	36.36%	25.45%
Los Angeles Lakers	2	1	33.33%	23.33%
San Antonio Spurs	11	2	15.38%	10.77%
Detroit Pistons		6	1	14.29%	10.00%
Phoenix Suns		7	1	12.50%	8.75%
Utah Jazz		15	1	6.25%	4.37%
New Orleans		15	1	6.25%	4.37%
Dallas Mavericks	25	1	3.85%	2.69%
Orlando Magic		30	1	3.23%	2.26%
Cleveland Cavaliers	30	1	3.23%	2.26%
Houston Rockets		35	1	2.78%	1.94%
Washington Wizards	50	1	1.96%	1.37%
Denver Nuggets		60	1	1.64%	1.15%
Tornto Raptors		100	1	0.99%	0.69%
Philadelphia 76ers	200	1	0.50%	0.35%
Atlanta Hawks		300	1	0.33%	0.23%

After scaling, the probabilities for the top teams are even lower. Even before scaling, obviously no team has a 50% or better odds (that would require 1-1, right?). But of course these are just betting odds, and one can argue these do not reflect the true probabilities.

 

[Stargazer]

Actually, I'm amazed at some of these odds:

1. Detroit's odds should be higher, simply because they (and Boston) are in the weaker east. They are much more likely to make the finals (they really only need to beat Boston), and so should have a higher chance of winning.

2. LA's odds should be lower, for the same reason. They will have to beat good teams just to get to the finals, but that really isn't reflected in the odds.

3. Most surprising, though, and most interesting, is that the Suns and Spurs together have nearly a 1 in 4 chance of winning it all -- yet they are playing each other in the first round. Actually, that seems right to me - it means that the winner of the series has a good shot to go all the way.

 

[elindholm]

Ok then, here it is:

Cool, much better.

After scaling, the probabilities for the top teams are even lower. Even before scaling, obviously no team has a 50% or better odds (that would require 1-1, right?).

Right. I agree that this year is unusually wide-open.

Maybe it would be interesting for some people to post what they think the actual probabilities are. In my opinion some of the longshots are way too high -- by which I mean, not longshot-y enough. Does anyone really think that there is almost a 4% chance that the champion will be one of the bottom five teams? I sure don't. I'd put their collective chances at zero (or 1/1000, if you insist) and be done with it.

 

[Irish]

The big difference between the NCAA tournament and the NBA playoffs is that the fluky one game explosion that kills so many college contenders is only one game of 7 in the NBA. For example, last year the Nuggets won a playoff game against the Spurs, but over their stats were very one sided to the Spurs.

 

[cobbler]

Utah at 15:1 is pretty nice.

Im loving my $100.00 bet over the offseason at 25:1 for the Lakers.