Public team bias

**EXhoosier10** · 02-01-13, 02:37 PM

Why limit this to just road games and why just spreads of just 3 points or less?

**RabidGolfer** · 02-01-13, 02:54 PM

Originally posted by EXhoosier10

Why limit this to just road games and why just spreads of just 3 points or less?

My thinking is threefold:

When the dog has home team advantage, the variance is likely bigger between the true line and the book line. Home team advantage seems to be amplified when it's David vs a public Goliath, and David actually is a good, yet underrated team with a few hidden stones in his sling. The Warriors this year have been a good example of a team like this.

Plus, I like the smaller spreads because theoretically that would indicate the dog is still a solid team, but just isn't getting the points because they don't have the fan-base.

Thirdly, the public is more likely to pile on to a favorite when the spread is -1 to -3, but probably would think twice if the spread is -7. In other words, I'm banking on the "mindless masses" ignoring the spread because it's so low and simply picking their team to win.

Bottom line is: If my first reaction when I see a line is, "wow, with that low spread it's a no brainer to pick the road favorite", then it's likely a good candidate for my "Home Dog" theory.

What do you all think? Am I on to something?

**RabidGolfer** · 02-01-13, 06:48 PM

I'm going to put this to the test on tonight's game in Indiana. Pacers +3

**LtDementia** · 02-01-13, 09:43 PM

Originally posted by RabidGolfer

I'm going to put this to the test on tonight's game in Indiana. Pacers +3

Well, you might be on to something.... In hindsight, you could have bet the ML on this one.

**RabidGolfer** · 02-01-13, 10:46 PM

Originally posted by LtDementia

Well, you might be on to something.... In hindsight, you could have bet the ML on this one.

Exactly. I've seen this happen too many times for it to be a coincidence. Back on 1/14/13 when the Wiz were still sucking, they were hosting Orlando and they were a -1 favorite. I remember the majority of this forum took Orlando +1, and I took the Wiz. I didn't know anything about either team, but I figured the line looked too good to be true considering the Wiz overall win/loss record, so I took them. Ended up 120-91 blowout for the Wiz. Apparently the stats with Wall returning to the team were impressive enough that the books were giving the Wizards credit, but the public sentiment against the Wiz was keeping the line down to -1 (probably should have been -10, cause they were just starting their hot streak). I'm not saying it's foolproof, but I'd like to back test it as I have a hunch it would be right over 60% of the time.

**ask** · 02-01-13, 11:57 PM

Did some quick analysis of your theory on the last 6 seasons. Not profitable. Around 120-150 games per season match the criteria of being favorites on the road within a spread of 3 points or less. Three of six seasons are below a total win percentage of 50, while the other three are above. But no season is profitable at -110.

**LtDementia** · 02-02-13, 12:50 AM

Originally posted by ask

Did some quick analysis of your theory on the last 6 seasons. Not profitable. Around 120-150 games per season match the criteria of being favorites on the road within a spread of 3 points or less. Three of six seasons are below a total win percentage of 50, while the other three are above. But no season is profitable at -110.

Any chance that they might be profitable wagering the ML?

**tto827** · 02-02-13, 12:52 AM

You are in essence attempting to bet a system. Systems will revert to 50% in the long run. You can profit short term, and like the stock market, you can make money if you get out at the right time.

**RabidGolfer** · 02-02-13, 01:26 AM

Originally posted by ask

Did some quick analysis of your theory on the last 6 seasons. Not profitable. Around 120-150 games per season match the criteria of being favorites on the road within a spread of 3 points or less. Three of six seasons are below a total win percentage of 50, while the other three are above. But no season is profitable at -110.

My theory is actually to fade the public team that's on the road When the spread is three points or less. Not sure if that's what you were talking about. Anyways thanks for running the analysis. I may go back and look at it based on just a few heavily public teams like the Lakers and the Heat To see if they are getting extra points. It makes sense that they would since a certain percentage of the public will always take them regardless.

**RabidGolfer** · 02-02-13, 01:28 AM

Originally posted by tto827

You are in essence attempting to bet a system. Systems will revert to 50% in the long run. You can profit short term, and like the stock market, you can make money if you get out at the right time.

That's true even in betting? I guess that makes sense if the books get wise to the system and negate it, or enough people start using it that it becomes diluted. That's too bad

**tto827** · 02-02-13, 01:37 AM

Originally posted by RabidGolfer

That's true even in betting? I guess that makes sense if the books get wise to the system and negate it, or enough people start using it that it becomes diluted. That's too bad

**byronbb** · 02-02-13, 01:03 PM

I am pretty confident you could do analysis of the line movement and make estimates at a line's high or low point.

**tto827** · 02-02-13, 02:13 PM

Originally posted by byronbb

I am pretty confident you could do analysis of the line movement and make estimates at a line's high or low point.

SBR Forum - Sports Betting Talk - Sportsbook Review Forum

http://forum.sbrforum.com/ncaa-basketball-betting/2154972-wave-theory-experiment.html

Sports betting and handicapping forum: discuss picks, odds, and predictions for upcoming games and results on latest bets.

Check out this thread.

**RabidGolfer** · 02-02-13, 03:38 PM

Originally posted by tto827

http://forum.sbrforum.com/ncaa-baske...xperiment.html
Check out this thread.

Nice, I like what he's doing with that.

**RabidGolfer** · 02-02-13, 05:06 PM

Okay, I just did a cursory analysis of this season based on the Lakers and the Heat. Both public teams.

Theory: Fade a public team on the road against a home dog with a low spread.

Criteria:
Home team is a dog between 0 and +3
Away team is a public favorite
Home team has had one or more days of rest
Away team has had no more than one day of rest

Miami Results: 1-5 ATS

2/1/2013	at INDIANA	-2	184.5	89-102	L	L
1/30/2013	at BROOKLYN	-2	191.5	105-85	W	W
1/14/2013	at UTAH	-2.5	198	97-104	L	L
1/8/2013	at INDIANA	-1.5	187.5	77-87	L	L
11/14/2012	at LA CLIPPERS	-1	197.5	100-107	L	L
11/11/2012	at MEMPHIS	-1.5	194.5	86-104	L	L

LA Lakers Results: 2-4 ATS

2/1/2013	at MINNESOTA	-2.5	200.5	111-100	W	W
1/30/2013	at PHOENIX	-3	204	86-92	L	L
1/21/2013	at CHICAGO	PK	194.5	83-95	L	L
12/4/2012	at HOUSTON	PK	215.5	105-107	L	L
11/24/2012	at DALLAS	-1.5	205	115-89	W	W
10/31/2012	at PORTLAND	-2	192	106-116	L	L

So gentlepeople, we may be on to something. Granted it's a small sample size, but so far fading a public team with the above criteria is yielding a 75% success rate.

What are other highly public favorites that I should analyze? Would OKC or LAC fall into this category? Ideally, I'm thinking of picking the 5 most popular teams for this analysis. Your input is appreciated.

**RabidGolfer** · 02-02-13, 05:10 PM

Actually, a more interesting point is the dog is 75% on the ML. That's some profit right there!

**gotov** · 02-02-13, 06:30 PM

Would u take pistons tomorrow if lakers were -3 or under at detroit?

**RabidGolfer** · 02-02-13, 06:51 PM

Originally posted by gotov

Would u take pistons tomorrow if lakers were -3 or under at detroit?

Yes, definitely. Especially considering the troubles the Lakers have been having.

**ask** · 02-02-13, 10:00 PM

Originally posted by RabidGolfer

My theory is actually to fade the public team that's on the road When the spread is three points or less. Not sure if that's what you were talking about. Anyways thanks for running the analysis. I may go back and look at it based on just a few heavily public teams like the Lakers and the Heat To see if they are getting extra points. It makes sense that they would since a certain percentage of the public will always take them regardless.

Aha. Understood it differently - as betting the road team ATS while being favorite and matching a spread of 3 points or less. Correct me if I'm wrong, but based on your other post #16, your approach is to bet against the public favorite road team while the spread is 3 points or less, the "... home team has had one or more days of rest" and the "... road/away team has had no more than one day of rest". My initial thought is, I guess there's not a lot of games per season that match those criteria (without saying anything about the success rate). I'd be more than happy to help doing some backtesting for you, just say the word.

**evo34** · 02-02-13, 11:52 PM

The system has not been good historically. Home dogs of +3 or less when getting < 25% of the # of bets have gone 133-140 ATS since 2005. In general, my feeling is that the NBA is the least friendly sport for technical analysis.

**RabidGolfer** · 02-03-13, 02:39 AM

Yes please guys, if you have the ability to do back testing by all means. Sounds like Evo may have done it already, but it needs to be limited to just a few public favorites. The reason being that, two bad teams aren't going to have a skewed line. But between a good public team and a good but less popular team, the public will take the popular team and the books will therefore skew the line to preserve some balance in the betting (theoretically).

If anybody does backtesting, please limit it to the Miami Heat, Lakers, Bulls, OKC, and Knicks. One or two more is fine, but I don't want to dilute the results.

**Miz** · 02-03-13, 09:52 AM

The wave theory stuff is interesting. For those without sound models yet, it is a lazy, yet possibly viable, approach. I may back test this in the next couple of weeks as time permits. It is really easy to test.

I imagine something like, setting a limit on the ATS streak for the two teams (opposing streak). Begin with a limit of 0 (all games) and subsequently increase it to 5 or 6 (where there are likely very few games that would qualify).

If it works, I promise to keep the results to myself, but if it doesn't I'll gladly post about it

Or maybe I'll just not post either way, and let the motivated among us do the research themselves.

The most obvious flaw in this is that there is no way to adjust for market changes. If you had a real model that gave you a true estimate for the line, then an ATS streak input may improve it, but it should be one of many inputs IMO.

Enjoy your day.

**ask** · 02-03-13, 09:56 AM

If we're about to do backtesting, there must be a difference in who's public teams in each of the previous seasons? Doing some backtesting on Lakers in the last 6 seasons, the number of games that match your criteria is < 5, while being < 10 without the 'rest' criteria.

**brettd** · 02-03-13, 10:21 AM

Originally posted by RabidGolfer

Okay, I just did a cursory analysis of this season based on the Lakers and the Heat. Both public teams.

Theory: Fade a public team on the road against a home dog with a low spread.

Criteria:
Home team is a dog between 0 and +3
Away team is a public favorite
Home team has had one or more days of rest
Away team has had no more than one day of rest

Miami Results: 1-5 ATS

2/1/2013	at INDIANA	-2	184.5	89-102	L	L
1/30/2013	at BROOKLYN	-2	191.5	105-85	W	W
1/14/2013	at UTAH	-2.5	198	97-104	L	L
1/8/2013	at INDIANA	-1.5	187.5	77-87	L	L
11/14/2012	at LA CLIPPERS	-1	197.5	100-107	L	L
11/11/2012	at MEMPHIS	-1.5	194.5	86-104	L	L

LA Lakers Results: 2-4 ATS

2/1/2013	at MINNESOTA	-2.5	200.5	111-100	W	W
1/30/2013	at PHOENIX	-3	204	86-92	L	L
1/21/2013	at CHICAGO	PK	194.5	83-95	L	L
12/4/2012	at HOUSTON	PK	215.5	105-107	L	L
11/24/2012	at DALLAS	-1.5	205	115-89	W	W
10/31/2012	at PORTLAND	-2	192	106-116	L	L

So gentlepeople, we may be on to something. Granted it's a small sample size, but so far fading a public team with the above criteria is yielding a 75% success rate.

What are other highly public favorites that I should analyze? Would OKC or LAC fall into this category? Ideally, I'm thinking of picking the 5 most popular teams for this analysis. Your input is appreciated.

What you are essentially doing is trying to identify lop sided action heuristically using rule trees. Instead off hoping you have identified a valid subset of all games that have lop sided action heuristically, just directly take a look at the book's percentage splits on a game.

And yes, there is a somewhat negative relationship (it's not linear though) between a percentage increase in the proportion of action down on one side of the spread, and the probability of the other side covering.

However this is not enough to beat the spread at better than >= 52.38 ATS% long term.

**RabidGolfer** · 02-03-13, 11:24 AM

Geez, I wish I understood half of what you just said...ha! Look, If I'm right there are two games that qualified today. Boston Vs clippers And Pistons vs Lakers. Both fit the the criteria, and both will probably win on the Moneyline.

**brettd** · 02-03-13, 12:37 PM

I should have clarified, I use SportsInsights to track the physical number (though they do not have access to the value of each bet) of bets on each side of the spread. This can act as a guideline to the weight of money on either side of the spread (identical to what you're trying to achieve heuristically).

Your idea won't work to make +EV, and has been tried the world over before. The utility of simple angles such as this were incorporated into the efficiency of the spread many many years ago.

**RabidGolfer** · 02-03-13, 03:04 PM

It would seem to me that In a game like today's Clippers at Celtics, the books would almost have to give the Celtics some extra points to avoid a potentially devastating loss. Lets say the properly handicapped line should be Boston -2 instead of Clippers -1. In that case 90% of the money would probably go to the Clippers due to publicity of the team, and prominence of the players, and their record. Lets say the Clippers happen to win ATS. The books would take a devastating loss having to pay out 90%. For their own safety they need to set the line to at least keep it 60/40 to cover their a$$es.

Lets look at it from a different perspective. Take a public team like the Heat. When have the Heat ever been a dog? Is it really true that they are ALWAYS favored to win statistically? Of course not. But the lines would indicate that are. So are the lines really that efficient, or are they set to benefit the book? My theory is that the lines are sometimes inefficient with these high profile teams because sometimes they really should be the dog, but the books would risk major losses if they set the line correctly.

**RabidGolfer** · 02-03-13, 03:43 PM

So far since starting this thread, All three games that have qualified have beat the spread, and two won SU.

**brettd** · 02-03-13, 07:31 PM

Originally posted by RabidGolfer

It would seem to me that In a game like today's Clippers at Celtics, the books would almost have to give the Celtics some extra points to avoid a potentially devastating loss. Lets say the properly handicapped line should be Boston -2 instead of Clippers -1. In that case 90% of the money would probably go to the Clippers due to publicity of the team, and prominence of the players, and their record. Lets say the Clippers happen to win ATS. The books would take a devastating loss having to pay out 90%. For their own safety they need to set the line to at least keep it 60/40 to cover their a$$es.

Lets look at it from a different perspective. Take a public team like the Heat. When have the Heat ever been a dog? Is it really true that they are ALWAYS favored to win statistically? Of course not. But the lines would indicate that are. So are the lines really that efficient, or are they set to benefit the book? My theory is that the lines are sometimes inefficient with these high profile teams because sometimes they really should be the dog, but the books would risk major losses if they set the line correctly.

The betting public rarely dictates the spread in the NBA. Dude trust me, there are only a few groups on this planet that beat the NBA at this time of the season. The groups that can, do so, because they have the resources to analyse the game into the depth that your typical handicapper could not achieve (video analysis, codification of custom game variables, access to early injury information, and very importantly, the quantification of the impact of injury).

The utility of most angles and methods are incorporated into the line. I've built all sorts of models against the NBA spread, and they go to shit at this time of the season; I've tried time series analysis of ATS history (ARIMA, and state-space). I've built your 'vanilla' regression and logistic regression models, neural networks, decision trees, all sorts of power ratings ('Massey' type vector ratings, ELO, MCMC, etc) and the list could go on.

Your method will not show +EV against the NBA spread. I would focus on props or try to find value in QTR time lines.

**EXhoosier10** · 02-04-13, 09:52 AM

Originally posted by RabidGolfer

So far since starting this thread, All three games that have qualified have beat the spread, and two won SU.

The stock I picked has gone up the past three days in a row ---> My system to pick stocks must be a winner. Same logic, same uselessness.

**RabidGolfer** · 02-04-13, 03:46 PM

Originally posted by EXhoosier10

The stock I picked has gone up the past three days in a row ---> My system to pick stocks must be a winner. Same logic, same uselessness.

I get that. Doesnt prove anything, but still supports the theory. Ignore my small sample size and think about the logic from the book's perspective. If you we're the book and you thought a line on the Heat should be +5, would you put it there? You would be crazy to because 99% of bettors would take those points in a heartbeat. If the heat happened to cover, you would be screwed. So they set the spread high to cover their own a$$.

If you think it's far fetched, how come they've only been a dog in two games this year? They're good, but they shouldn't always be the favorite. It's because the books are trying to set the lines where the public will bet, and not necessarily where the game will be.

If I'm wrong and the books are always trying to give sharp lines vs always trying to maximize profit, then explain to me what their motivation would be.

**RabidGolfer** · 02-04-13, 03:50 PM

Originally posted by brettd

The betting public rarely dictates the spread in the NBA. Dude trust me, there are only a few groups on this planet that beat the NBA at this time of the season. The groups that can, do so, because they have the resources to analyse the game into the depth that your typical handicapper could not achieve (video analysis, codification of custom game variables, access to early injury information, and very importantly, the quantification of the impact of injury).

The utility of most angles and methods are incorporated into the line. I've built all sorts of models against the NBA spread, and they go to shit at this time of the season; I've tried time series analysis of ATS history (ARIMA, and state-space). I've built your 'vanilla' regression and logistic regression models, neural networks, decision trees, all sorts of power ratings ('Massey' type vector ratings, ELO, MCMC, etc) and the list could go on.

Your method will not show +EV against the NBA spread. I would focus on props or try to find value in QTR time lines.

I'll take your word on this (again, you lost me at ARIMA). Guess I should stick with the day job.

**tto827** · 02-05-13, 03:53 AM

Originally posted by RabidGolfer

I get that. Doesnt prove anything, but still supports the theory. Ignore my small sample size and think about the logic from the book's perspective. If you we're the book and you thought a line on the Heat should be +5, would you put it there? You would be crazy to because 99% of bettors would take those points in a heartbeat. If the heat happened to cover, you would be screwed. So they set the spread high to cover their own a$$.

If you think it's far fetched, how come they've only been a dog in two games this year? They're good, but they shouldn't always be the favorite. It's because the books are trying to set the lines where the public will bet, and not necessarily where the game will be.

If I'm wrong and the books are always trying to give sharp lines vs always trying to maximize profit, then explain to me what their motivation would be.

I am with you on this thought. A books goal is to get 50/50 action, period, end of discussion. That's why I believe it is possible to beat the books simply by being "sharper" (for lack of a better term) than the public. If books set lines that were the true expected outcome, then they would be nearly impossible to beat.

Here is a challenge. You cannot bet games with major injuries. See if you do better against the books, or against the KenPom projection. Perfect example of this would be Duke at Florida St last weekend. Duke has struggled on the road this year (especially ATS) and was only -6 ATS and -250 or so on the ML. Florida St is known for good D and playing well at home. Ken had Duke winning this game 80%+ (and by double digits I believe), and Duke rolls easily.

Actually as I think about it, bad example as Ken doesn't account for injuries, so that could explain some of the line discrepancy, but I wasted my time writing it so.

**EXhoosier10** · 02-05-13, 10:26 AM

Originally posted by tto827

I am with you on this thought. A books goal is to get 50/50 action, period, end of discussion. That's why I believe it is possible to beat the books simply by being "sharper" (for lack of a better term) than the public. If books set lines that were the true expected outcome, then they would be nearly impossible to beat.

Here is a challenge. You cannot bet games with major injuries. See if you do better against the books, or against the KenPom projection. Perfect example of this would be Duke at Florida St last weekend. Duke has struggled on the road this year (especially ATS) and was only -6 ATS and -250 or so on the ML. Florida St is known for good D and playing well at home. Ken had Duke winning this game 80%+ (and by double digits I believe), and Duke rolls easily.

Actually as I think about it, bad example as Ken doesn't account for injuries, so that could explain some of the line discrepancy, but I wasted my time writing it so.

Now that you bring up Pomeroy, I've been tracking his projections for the last few days and probably 80-90% of hist projections are within 2.5 points (as I say this, 8 of the 17 games today are 'off' by more than 2.5 points, so it could be closer to a 75/25% split where vegas is pretty much spot on). I'm not sure if there are "public" teams in NCAA before the tournament begins (as I would imagine there are not too many "squares" who bet college without knowing much about it like pro sports, but i could be wrong), so maybe line setting for NCAAB regular season is trying to split 50/50 and therefore wouldn't even apply to your idea.

With that being said, Golfer, what is enough of an edge in terms of points per game where you would foresee being profitable long term? If a fair line should theoretically be -3 and you get it at -1 every game, is that going to be enough of a margin to win decent money long term, or are you more likely to hit 53 or 54% and scrape out small profits since your subset of games is so small?

While 54% is winning, I was +3% ROI in MLB this season (which would mean over 500 bets I had an average of 3% edge per bet). If you focus on those 3-5 'public teams' and only bet away games where they're < 5 point favorites, you're looking at (200 away games for 5 teams, maybe 30% are within 5 point favorites?) 60 games over a season. So even if this was a successful system, you'll need more games to actually make a decent return without risking a big chunk of your br per bet.