Some tech angles week 1

And now a stats based discussion...none of this malarkey on websites with no actual stats or evidence to prove what they are saying is true...."shop for line value", "don't bet over your  head", "fade the public"...blah, blah, blah....I never see any hard statistics to say what happens by waiting to bet later in the week, or what the percentages are by going against the public....it is a bunch of filler of non-useful information to get people to come onto someone's site, with nothing useful to refute or back up the commonly held wisdom.

You may have read that turnovers and yards per pass attempt are the two most crucial statistics in winning football games.

True or not true?

We'll see how a team with the greater  net yards per pass does, and MOST IMPORTANTLY, "can I win using this information?"

Net yards/pass means we are taking the offensive yards gained per pass and subtracting the amount of yards allowed per pass on defense.  So if team A throws for 8 yards/pass and allowed 7 yards/pass in a game that is +1.  Obviously those stats are reversed for the team they played so that their opponent team B had a -1 yards/pass for the game.

The team with the net advantage wins the game 73% of the time and covers the spread 69.5% of the time, which, no one ever hits that percentage, so what gives? 

The 64,000 dollar question is...."DO WE MAKE MONEY BY TAKING THE SUPERIOR NET PASSING TEAM IN NFL BETTING based on past performance?"

Let's run a test with 31 years of NFL data in it, and we'll answer that question.

Here is the query text that I use in sportsdatabase.com

tA(YPPA-o:YPPA)-oA(o:YPPA-o:YPPA)>0 and game number<16

"And the answer is".......(drum roll!!)

"No, we don't make money by taking the superior net passing team."

The results we get are 3199-3312 ATS, which for you that are bad in math, is 113 games below 500, or 49.1%.

Disregarding the defensive component and just comparing past offensive passing yards/attempt yields 49.8% ATS covering percentage, a little better but still south of the Mendoza line.

In actuality, our winning percentage is higher by fading the superior passing team, but still not profitable....or to use another term, we expect regression to the mean. 

Sportsbooks hire people with powerful computers that have figured out the importance of yards per pass...to beat the bookie there will have to be other outside-of-the-box kinds of thinking, that perhaps they aren't aware of. 

Meanwhile, sportsbooks and casinos continue to build taj mahal like structures financed by sports bettors and one armed bandit users.

How about turnovers then?....what happens against the spread with turnovers, AND, most importantly, can we make money knowing the history of two teams' past turnover history?

Teams that lose the turnover battle will lose the game 78.8% of the time and cover the line only 22.5% of the time....so of course we'd love to know which team is more likely to have more turnovers...THAT could lead us to having our own tropical island with beautiful women feeding us grapes and fanning us, so let's find out.

We again are going to enter in net turnovers per game, so if our good defense intercepts a couple of passes a game on average but our intellectually challenged quarterback fumbles once and throws an average of two interceptions a game, we'll be a bad team more than likely.

In sportsdatabase, TOM= turnover margin, and the higher more positive number means our offense is commiting more turnovers than the defense is getting....in other words, the more negative our turnover margin number, the better off we should be.

tA(TOM)-oA(TOM)<0

means we have the turnover advantage based on past average game statistics.

The results of our query gives us an against the spread record of 3324-3356 ATS, 48.5%, so hopefully you didn't make a down payment to that island salesman.

If we up that advantage to more than one turnover per game difference between the two teams, the results are WORSE....47.9%.

So, in both cases of "the most important statistics" that we've read about, the sportsbook linemaker also got the memo and has adjusted the line to take away any perceived advantage... IN BOTH CASES WE WOULD DO BETTER TO FADE PAST STATISTICS.

In fact in a game where a team has both the yards per pass AND the turnover margin disadvantage we cover at 52.4%, which allows us to eat bread and water for our sustenance, but not much else...still crummy, but soooo much better than the 47ish% we would cover by taking the team with the passing and turnover advantages.

Food for thought....

Then on the other hand....if we take the team with the net turnover/game advantage and the net passing yards/pass per game advantage in the PLAYOFFS ("Playoffs?!!"), those teams are 59-40 ATS.

In game 1s, how does a divisional underdog do that lost both games to that team the season previous?

D and game number = 1 and P:L and PP:L and DIV

48-28 ATS......Redskins, Bucs, Cardinals

If our divisional dog with at least double revenge didn't make the playoffs last season this moves to 48-26 ATS, and if their divisional opponent did make the playoffs, it has been 23-8 ATS

D and game number = 1 and P:L and PP:L and DIV and tpS(playoffs)=0 and opS(playoffs)>0

If their game one opponent won their last season's game 1 this moves to 16-2 ATS

Cardinals, Redskins and Bucs

Thanks, Mr. Russo!

Thanks, Mr. Russo and LordBennington!

Indigo, this is GREAT INFO! Much appreciated!

And now, we'll have a discussion on the bookmaking industry.  Obviously most of the USA has gotten out of the stone age and the lightbulb has come on that sports betting will always go on, and why shouldn't the individual states receive tax revenue for what is going to happen anyway with most money going off-shore.....a very similar situation to what happened with alcohol consumption in the prohibition days.

Some states have limited bookmakers, which is never a good thing...they would be much better served to have many options in one state, as competition drives better service and better prices.  Obviously, talking to actual sports bettors probably never happened for the decision makers in each individual states to the detriment to consumers.

Overseas, not sanctioned books will have to offer better prices to compete with the safety and guaranteed money that a local sportsbook offers.  However, the days of -110 against the spread betting may be numbered.  Outside of the USA, exchanges are making headway and they will be the way of the future, though traditional sportsbooks with their considerable deep pockets  will bribe all the important decision makers to keep the status quo.

For those of you that don't know the difference, traditional bookmakers offer -110, either way betting....if team A is playing team B, you can bet A or B and pay a 10% commision for the privilege of betting with that sportsbook.  Though you'll hear that sportsbooks want equal action to guarantee themselves a profit, in actuality the bigger sportsbooks deliberately accept risk on unequal action on  individual games.  They in advance know how the public will bet and where the most action will be.  The flaw in the system is that are expected to make money and there is risk in their business that unequal action can cause them to lose.  It rarely, rarely happens that over the course of time that they do lose, as every year Las Vegas churns out a net profit of 5% of total wages.  That comes with subtracting the expenses of running their operation, so there actual gross profit is higher than 5% after paying off wages, etc.  

Betfair, an exchange, did a study of their 80,000 customers....3,000 of them were long-term winners, which equates to 3.75% of the total betting population that they have.  That percentage  is very similar to what we hear, that not many make money in the long-term as sports bettors.

Some traditional bookmakers will lose sight of that fact, and will kick winners out of their books.  When 96+% of your bettors are losers, rather than kick out the winners, if they are intelligent, the concentration should be on generating volume, as pinnacle has done for many years, which enabled them to become the most successful offshore bookmaker.  They never cared about kicking out winners, and in metaphysical parlance, their karma was good, as was their business model.

Exchanges take the risk out of bookmaking completely for the bookmaker.  If someone wants to bet team A, they go to their  sports betting exchange site, they see if the -6 that our customer wants to bet is available and he bets it, and he pays a commission if he wins, which with matchbook has been 2% and for other exchanges ranges from 3.5%-5%.  If his price is not available, he makes an offer, which, if it is in line with what is commonly available will almost surely be covered, unless he bets 3 minutes before game time, and then it is possible that his offer will not be filled. 

The bookmaker has guaranteed  balanced action on both sides, so the house has absolutely zero risk.  And, obviously because of this, the bookmaker could care less who and how much anyone wins.  So, no one gets kicked out for being a winner...it is similar to pari-mutuel horse racing where the house is guaranteed a certain percentage of profit.  When Bill Benter and Alan Woods were making 100s of millions of dollars on horse racing in Hong Kong, it was not big deal to those running the horse racing betting in HK.

In Australia, betfair lists the amount of money wagered on anyone game.  Australia typically hates undercutting or competition, and their commission is 5%, which I am sure the powers-that-be mandated or else the traditional sportsbook would cry "unfair" which roughly equivalent to paying -110 on a spread bet...sometimes the amount bet on one game is over a million dollars, so betfair takes home 50,000 bucks guaranteed on that game.  And surprisely, the most money bet is on in-game wagering.

So, is it unfair, or is in the Sam Walton paradigm that you offer awesome, better-than-the-competition prices and make up for the lessened profit margin per bet by increasing volume, business model.  Seems to have worked for Sam Walton and Wal-mart, I'd say.

In the rest of our existence you go online to shop or to your local store and one vendor is offering a  price of 1500 bucks for a laptop and another offers the exact same thing for 1400.  You buy the one that costs 1,400 bucks, end of story.  

It shouldn't be so different in bookmaking, as competition should drive prices down.

There will come the day when some HUGE player, like the US government, or google, or Apple or someone of that ilk becomes a player in the USA betting market offering an exchange, at reduced prices.  It is an awesome business model with no risk to that company and it will change the sports betting landscape forever.

I got the math wrong for the betfair model.

If a million dollars is bet, in the betting exchange scenario there  would have  a guaranteed wager cash of 500,000 on one side and a subsequent matching  500,000 on the other.....they collect 500, 000 on the losing bets, and take out 5% of the 500,000 dollars in winning bets, which is 25,000 dollars commission per million dollars bet.

In a traditional book, there would not be a perfect 50-50 division of cash on either side of a bet and the book assumes the risk of uneven action, making sportsbook managers and linemakers liable of losing their jobs if they get it wrong consistently...however for theoretical purposes, if they did have  a game in which a million dollars was bet, 1/2 of that equally on each side.....500,000 would be kept as losing wagers, and 500,000 dollars minus their commission of 45,000 would be paid out to winning punters.

So the difference is between the two is 20,000 dollars in commission per million dollars bet (45,000 versus 25,000), that the bookie gets and you don't.

One problem with exchanges though...NFL markets aren't loaded with liquidity. For now that's a big negative.

I like the Pittsburgh Steelers to win more than 9.5 games...they have a weak and low miles travel schedule. Their D is arguably top 5 and with Ben back and healthy with some new weapons on offense, they might easily be 12-4. 

The toughest teams they face are Baltimore twice, the Titans, the Colts, Texans, Eagles, Bills, Cowboys...

This thread is like an advanced course in football handicapping.

Low liquidity?  Betfair has 1-3 million matched side bet placed on each NFL l game....population of Australia is 25 million, so I suppose your definition of low liquidity and my definition don't agree.  Liquidity obviously would vary according to the population of the offering.  If we had a nationwide exchange like the New York Stock Exchange the argument is a moot point.  If we had an individual state with an exchange with a population of a couple million people perhaps that is an argument.

The issue for exchanges would be how to handle teasers and parlays...those are not offered presently on the exchange model, and probably wouldn't be.  THOSE would be low liquidity propositions.

Someone with direct knowledge of the betting industry could tell us the percentage of profit from sportsbooks comes from the exotic wagers of teasers and parlays, that would be interesting.

I don't think the sportsbook model we have now will ever go away, but so far there has not been any competition based on price, which has happened offshore, with Heritage, 5dimes and Pinnacle offering reduced juice...maybe there's others, that I am not aware of.  A Sam Walton type innovator perhaps will do that at some time, if he is actually allowed to by regulators.

The practice of kicking winners out should be regulated out of existence, which an exchange would do.  It is a ridiculous practice when a sportsbook holds a 5% advantage and only 3% of the bettors win.  We'll see what the future holds.

Here is a sportsbook story.  

About 20 years ago, there was a sportsbook, Aces Gold....Aces Gold lapped the field with their offerings.  On Friday nights, they offered JUICE-FREE wagering on NFL games during the football season.  (Just goes to show a book could make money WITHOUT charging any juice.). It was brilliant, because the management saw the NFL is much more difficult to beat than college football and it attracted funds away from the more beatable NCAA football games the next day. 

This went on for a few years and obviously it attracted the sharp money who got a greatly reduced price.  

For everyone's awesome strength there is a equal weakness....rather than just count the cash and buy up half of the city where Aces Gold was located, the owner made a fatal error of wanting to gamble.

He offered juice free wagering on the Super Bowl, between the Rams and Patriots in 2001, the game where it subsequently came out that Mr. Belichick hired spies to look into the Rams' Super Bowl practices to get their plays and audibles.  The consensus line on the Rams was -14. If I remember correctly Aces Gold offered the Patriots at +14.5 juice free, which of course  attracted a TON of Patriots money.

From the stories put out, he had 14 million in excess unbalanced action on the Patriots.  Normal protocol would've been to shift his line or to lay off to other sportsbooks the unbalanced action, but no, he obviously wanted to take the Rams.

Why Mister Bookmaker?...you had a good thing going!

We all know what happened, the Belichicks pulled the big upset and the Rams players, who had one of the best offenses in pro football history after the game said,...."it's as if they knew our plays!!"

Aces Gold and the owner disappeared into the night, never to be seen or heard from again.

Huh...mmmm...interesting. Do you know if the liquidity rises a lot in the last 8 hours before the match? Problem is that it's night here when NFL matches are live. So my bets are most of the time 8 hours in advance (or earlier).

Post #24 correction

Should read: 

Teams that made the playoffs last season on Monday night, game 1s  have been 5-2 ATS (though if their game 1 opponent DIDN'T make the playoffs this is only 1-1).

Titans 

All other days/nights playoff teams from previous season in game 1s as away dogs have been 12-33 ATS (6-19 ATS if their game 1 opponent also made the playoffs.)

tpS(playoffs)>0 and AD and game number=1 and day!=Monday

AGAINST Packers, Texans

thx for info in this thread 

What about days of the week?...what days are dog days and what days are NOT good days historically to bet underdogs?

For the first two months of the season, September and October.....  Sunday days and Monday nights are the best days to bet underdogs.

Sunday day dogs have hit close to 52%, and Monday night dogs have hit close to 54% for the past 31 years, first two months of the season...after that into November Monday dogs become a pathetic play.

Early season Thursday night underdogs....38-42 ATS

Early season Sunday night underdogs 96-116 ATS

Early season Monday night underdogs  136-120 ATS

What about if there's no fans in the stands?....unfortunately we have nothing in our historical database to compare.

Continuing on with a discussion of days of the week.

"Ah come on!...that doesn't really matter does it?"

Yeah, it does.

If you remember anything out of these threads, it should be that we expect regression on Sundays (and Saturdays) and favor the better team on Thursdays and Mondays.  (In an earlier thread I mentioned Mondays as a dog night, so that is something that I will have to look into.)

I looked at 31 years of data and looked at how teams did the season prior, and, out of all the data subsets, the best performing teams won 2 or less, but not greater than 9 games less the year prior performed.  So our play ON team won 5 games last season and their opponent won at least 7 but less than 15 last season, our play on team qualified.

Here's the stats.

Home   671-640  ATS, 51.2%

Away   1086-951 ATS, 53.3%

Neutral 9-13 ATS 

And here are the days of the week for all dogs that fit our criteria

-10<PRSW - o:PRSW<-1 and D and game number < 16 and site!=neutral and day

Thursday 48-75 ATS

Monday 108-111 ATS

Saturday 30-19 ATS

Sunday night 101-86 ATS

Sunday day 1468-1300 ATS

By time zone and eliminating Saturday and Thursday games...our dogs that qualify

Pacific  227-228 ATS, 49.9%

Eastern 871-768 ATS, 53.1%

Central 386-316 ATS, 55.0%

Mountain 117-93 ATS, 55.7%

Why do the Eastern and Pacific time zones perform worse?....your guess is probably better than mine...maybe it's because ever since you had that creek outside the back of your house changed and you are now moving your grass in a clockwise motion, whereas your father always mowed counterclockwise, and it affected air currents, which has caused an increase in sunspot activity over the western and eastern coasts of the USA, going back 31 years, which affected brainwave activity of all underdogs playing in New York and Los Angeles, but only on Sunday and Saturdays.

Yeah, I got it figured out.

Now, if you have some secret sauce statistic that works for you...add that onto this data and you should see increased results.

If you are taking dogs on Thursday and/or Monday nights, you are swimming upstream....we are expecting regression on the weekend games from what happened last year and taking the superior team on Thursdays at least is a good strategy...and Monday night is looking neutral.

Sounds super interesting, but I don't understand this:

I looked at 31 years of data and looked at how teams did the season prior, and, out of all the data subsets, the best performing teams won 2 or less, but not greater than 9 games less the year prior performed. So our play ON team won 5 games last season and their opponent won at least 7 but less than 15 last season, our play on team qualified.

Since 2012, a team that scored more points than it allowed the season prior and won less than 8 games has gone 7-2-1 against their regular season wins line.....Bucs

Since the start of the 2001 season when every team in the NFL started playing 6 divisional games in a season, there have been 24 teams that have won all 6 of their divisional games.

How did that team do the following season in divisional games?

Only one team, the 2014 Colts, won all of their divisional games the following season.  The mean for those 24 teams was that they won 4 out of 6 divisional games the following season, covering 3 out of 6 games.

Only one of those 24 teams, the 2011 Raiders, DID NOT make the playoffs the following season.

If they come a divisional dog in September or October they've been 6-3 ATS.....Chiefs, Packers

Conversely, there have been 30 teams that went 0-6 in divisional play the previous season,...how did THOSE teams do the following year in the division?

They won an mean average of 3 out of their 6 games the following season, also covering 3 out of 6.

They were particularly good to start the season going 23-12 ATS against divisional rivals in September......Chargers, Redskins, Lions

Hi DocS,

I will give examples.

Our team, team A, won 5 games last season, to qualify in my query, their opponent would have had to have won 7-14 games.

If instead, team A won 6 games, their opponent would have had to have won 8-15 games their prior season.

If instead, team A won 7 games last season, their opponent would have had to have won 9-16 games

Their present  opponent won 2,3,4,5, 6, 7, 8, or 9 more games the previous season than our team A did.

A game one non-divisional away favorite that won less than 8 regular season games last season

AF and game number=1  and not DIV and PRSW<8

5-15  ATS (-4.15)

7-12-1 straight up (-0.60)

5-15 over/under (-5.78)

VERSUS Raiders, Chargers