The sports betting industry is one that lacks regulation, meaning anyone can start a company, website or used car salesman persona to start selling picks. Because there is no regulation, handicappers can tout false records and promises of unimaginable wealth in order to obtain business.

While there are many legitimate and transparent handicappers in the industry, there are also an overwhelming amount who use fake names, flashy cars, women of questionable clothing and morals (we’re guessing) and unachievable records to convince new or uneducated bettors to buy their picks.

While this may sound a bit over the top, we often get calls asking why we don’t hit 70% of our games like many of the other services out there. Our response is always that a 70% win rate isn’t attainable over the long-haul.

To explain this in more detail, we analyzed the probability that a sports bettor can win 70% of all wagers to illustrate just how unrealistic this is.

For the purposes of this article, we chose the z-ratio (also known as z-score) to show how many standard deviations away from “expected” an event is.

**Example 1: No Edge**

This example assumes a handicapper who historically hits 50% of his games, meaning the handicapper does not have any edge when picking games. The data assumes 1,000 plays against the spread (with a vig of -110) over a calendar year, across all major US sports.

Desired Win Rate | Odds of Win Rate | Probability of Win Rate | Z-Ratio |
---|---|---|---|

50.0% | 1 in 2 | 50% | 0.0 |

51.6% | 1 in 6.3 | 15.9% | 1.0 |

53.2% | 1 in 44 | 2.3% | 2.0 |

54.8% | 1 in 700 | 0.1% | 3.0 |

70.0% | < 1 in a trillion | < .0000000001% | > 10.0 |

As you can see, a sports bettor with no edge has only a 2.3% chance of winning 53.2% of his games, which is just above the break-even point of 52.4%.

That same bettor has **less than a one in a trillion chance of hitting 70% of his games** over the course of 1,000 plays.

**Example 2: Good Handicapper**

We ran the same analysis above but this time assumed a skilled handicapper who historically hits 55% of his games. The data assumes 1,000 plays against the spread over a calendar year, across all major US sports.

For the record, we do believe there are good handicappers out there that are able to achieve 55% over the long-term, which is a very good win rate.

However, even in the case of a handicapper with a long-term expected winning percentage of 55%, a 70% win rate over a whole season (with 1,000 plays) would still be a hugely unexpected event.

In fact, this is still an almost-impossible “9 standard deviation event” with **the odds of a 70% or better win rate occurring less than one in a billion (.0000001%).**

Finally, an honest voice out there –

I’ve been betting sports since I moved to Vegas in 1980.

I’ve heard every tout and every gambler boast about how well he does picking winners against the spread. Unfortunately, both groups live in a world of denial and delusional rationalization.

What’s funny is gamblers and touts think they can fool me because they claim to be hitting 75% of their plays, but forget to tell me it was only on a Tuesday between 4 and 7 pm by winning 3 of 4 games. Complete nonsense for the long haul!

Good blog! I really love how it is easy on my eyes and the data are well written. I am wondering how I could be notified whenever a new post has been made. I have subscribed to your feed which must do the trick! Have a nice day! bdaebaacgdbd

Ok, but if you hit 70% then 70% is 0 standard deviations.

Exactly!

How do you calculate the “probability of win rate”? Am I missing something?

They didn’t just pull the 2.3% chance out of there ass, I am assuming.

I agree with the premise of the article, just trying understand how they got the 2.3% chance to hi5 52.3% of the time number

I disagree, i have been handicapping for 11 years now, i have record of every bet ever made, and In NcaaF alone, i have had 62% my first season and after that every year ranged in between 67%-78% i bet from 10-30 plays per week every week in ncaff season … now i dont do that well in other sports, NFL 55% NCAAB 59% MLB & NBA are losing props for me so i have stopped wagering on them 3 years ago… i bet spread only and totals… now im not the best in the world but what i am stating is if i can beat 70% in college football for the entire season im sure there is people better than me that can do it for a whole year, especially if they have a friend or 2 helping them research and crunch numbers…

just a thought

Do you post your plays publicly? If so where?

Thanks

Hi, you can see all of our Best Bet records here:

https://www.sportsinsights.com/sports-betting-systems/mlb-betting-systems/

Would like to talk about your college football records. If interested I can be emailed at tommydin924@gmail.com

Brian, you’re completely full of it, you know that? Nobody can win 67% – 78% of their wagers in football over the long haul, especially given the number of games you claim to bet on each week.

If you really had that ability, you could move to Las Vegas, place your bets at the sports books, and win millions of dollars each year. In other words, you’d even be better at handicapping than Billy Walters, which is almost impossible.

“Im sure there is better people than me”.

Tells you all you need to know about this cat.

I’ve been betting sports for 35PLUS years and have literally wagered MILLIONS! These guys are deadon!!! Sure, there may be an occasional, very,very rare exception, but just what are those odds??? If you want to have a fighting chance in hell of actually winning in the long run, LISTEN TO BRAD AND THESE GUYS!!! THIS IS YOUR GOLDEN OPPORTUNITY! SINCERELY Don (Been there and done it!!!)

Your article sort if glosses over your math and how you ran your regression. I know that might be a bit heavy for some people from a statistical explanation point I find this to be little light on details and proof.

PJ is simply using the normal approximation to the binomial distribution. You can find the relevant formula in any introductory textbook on probability. I’d type in here for you, if there were mathematical symbols on my keyboard.

All you guys got it wrong. Stats can be skewed any way you want them to look so just look at the black and white. All that matters is performance, right? In sports betting all one needs is a 52.4% or 53% to be profitable (taking into account for -110 vigorish or eleven divided by 21). Now whe none of you seem to take into account is this math dowss not account for ‘cherry picking’. There are certainly better picks than others. If one limits oneself to cherry picking the highest odds of willing games based on probability as a function of information it’s definately possible to pick 70%. What it comes down to is endurance and stamina keeping away from the distractions of Las Vegas. Simply put, your 1 trillion to one statements are raw math not taking into account of intelligence and everchanging information which ultimately contributes to the outcome of each pick.

I agree that over the long run (1000+ bets) it would be highly unusual for anyone to win or lose at a rate of 70% and that given Steve Stevens record of 1-4 (counting Pirelli’s recommendation that he says he got from Stevens) it is especially unlikely that Stevens has won 70%+ every year in the past . I say that not because it is true mathematically, or can be proven true mathematically, but because in 40 years of sports betting, twenty of them as a professional, I have never seen anyone even come close over 1000+ games.

Mathematically, however, it is not impossible. From a probability standpoint, every individual combination of wins and losses is equally likely to occur even in random results. Thus, flipping a coin you are equally likely to have heads come up 1000 times in a row as you are to0 have heads and tails alternate perfectly for the entire 1000 flips, or to get any other predicted combination of heads and tails. Of course the odds against any particularly combination appearing at random are immense ( 2 multiplied by itself 999 times) Try it on your calculator and see how far you get before your calculator goes on tilt.

Nevertheless, some combination must come up every time someone gets 1000 betting results, and nothing makes the result they got more likely than 70% wins. In addition, it is much easier to get 70% wins than it is to get 1000 wins or even to get the result any of you actually got in your last 1000 games. The reason is that only one result will satisfy the requirement of any set combination. To win 70%, however, you don’t need any given 70% combination. There are thousands of combinations that will provide 700 wins and 300 losses. To figure out the odds of ending up with a 70% record in 1000 bets, you must first multiply 2 by itself 999 times. Then, calculate all the possible combinations that will provide 700 W’s and 300 L’s. Then divide the first number by the second number.

But there is yeat another fly in the soup. We don’t even need exactly 700 wins and 300 losses. A result of 701 wins and 299 losses will do. In fact, any result that provides more than 700 wins and even a few results sli8ghtly less than 700 wins will do. Certainly, if Stevens won between 695 and 699 games the upward rounding wiould still get him hojnestly to 70% and to say he wasn’t 70% would be nitpicking. That means we can add hundreds of thousands of possible combinations in the 1000 games that will get him to where he needs to be.

Thus, the question, is: How did you calculate any of the probabilities you provide in the article? Do you have some formula that does it for you? A computer program perhaps? A little transparency is necessary if you want to credibly prove something scientifically or mathematically. Without transparency, you are simply spouting propaganda and become as bad as muck you claim to be raking.

Also, don’t forget, there is no reason to assume that the results will be completely random. Not every team is equally likely to cover the spread. Much has to do with the ability of the line maker to be accurate,in predicting the game, and the line maker is not trying to pick the game accurately — he is trying to even out public sentiment, which is quite a different matter. Then the public comes in and moves the line further to correct for the errors of the line maker in predicting sentiment. It is those errors in the line as a game predictor that account for the fact that year in and year out the Vegas casinos fail to hold their full 4.5% of the drop in sprots betting, and that the numbers indicate that the general public consistently wins between 51% and 52% of the time, and sometimes as much as 53%, with all bookmaker profits coming from sucker bets such as teasers, parlays, point buying, and parlay cards.

If we assume that a win is more likely that a loss, the odds against a final result providing between 700 and 1000 wins in 1000 diminish further.

Finally, you are engaging in circular reasoning. You try to prove that nobody can win 70% in 1000 sports bets by showing that it is unlikely for it to happen to someone who can’t win 70%. In other words, you are providing a self-fulfilling prophesy. What have you done to show that anyone claiming to win 70% has not actually done so? What sport are using for the results? How good is the line maker? How much has the line been skewed by public? What is the probability that any game will cover the given spread against any other team. I can prove to you all day long that a 70% result is highly likely depending on the betting propostion compared to the odds. The traveling con men used to win 70% of the time from the uneducated frontier suckers who made the incorrect assumption that the number 3 and the number 7 were equally likely to appear on a dice toss. If I bet laying 11-10 that a star NFL kicker will make a filed goal, I will certainly be a big 70% plus winner. If any handicapper waits and only bets when he can take advantage of some huge edge such as a star player getting hurt in the pregsame shoot around before the bookmakers get wind of it, he can most likely win close to 70% with nothing more. If you bet Withita State on College Baskets on the money line this season, you won 100%, and very few of those money lines fully reflected the very high odds that Wchita State would win straight up. You didn’t hear Stevens say the sport in which he hit 70% did you? He didn’t mention if that record was in money line games or in point spread games.

You make the invalid assumption that every game had a true 50% probability and that the handicapper had no advantage at all and could only pick games at random probability of 50%, and then told us that ending up with exactly a 70% ressult (as opposed to 70.4% or 71.2% is a one in a trillion shot. In other words you told us that someone who cannot win 70% is highly unlikely to win 70%. To which i say, “Duh!” You haven’t proven that it is highly unlikely to win 70% of ones bets for someone skilled enough to do it and who has historically dne it. It is like telling me that a golfer who averages 150 strokes for 9 holes is unlikely to hit a hole in one. What’s that got to do with Tiger Woods?

There are many things that cannot be proven mathematically. In fact, mathematically people can an will win between 70 and 100 games in 1000 even by just flipping a coin on spread games. That Stevens can’t or his unlikely to do it can only be proven anecdotally. That is we don’t see it everyday or any day for that matter. We can set a coputer to make random picks and note that we don’t see 700 wins in several million pretend bets. We accep0t fingerprints as a means of identification even though the laws of probability tell us that there can and will be multiple people with the same fingerprints. We accept fingerprints, simply because duplication is not something we can find in millions of examinations.

If you want ot discuss Steve Stevens as a fraud, best to stik to things like the lack of reason for doubling up bets, the mathematically proven Kelly criterion adn the proper amount to bet on a $50,000 bankroll with a 70% win probability and 11-10 odds, and the fact that he tells clients he is great money-management adviser, and he keeps repeating that he is there to help you stay in control aqnd then tells clients to make $99,000 in bets based on a $50,000 bankroll. Speak of the probability that a 70% + handicapper will have a 1-4 result (counting the game given to the Pirelli client) . Talk about his ttatement after the first loss that only the long run matters, and then doubling up into losses for no more reason than that there were two losses in the short term. Finally, criticize CNBC for allowing that con man to sit there are say, “I lost two and won one and my client made money. Don’t tell me about percentages,” while failing to disclose that the deal was a 50% commission on each winning game. the final bet was $66000 and the client won $60,000. Stevens, happy as a con man in Vegas, proclaims “Pay me.” The commission is $30,000. The prior loss is $33,000. the profit after the last bet is just $27,000, and the client needs to p;ay Stevens $3000 out of his own pocket on a pay only after you win deal. Final Score: Client: MINUS $3000. Steve Stevens: +$30,000 which, if his client had any brains at all, Stevens had to pay on medical bills to repair his bromen jaw and nose sustained just after he asked for the full win plus $3000 out of pocket for all that wonderful advice. It’s time for a letter-writing campaign to CNBC. then again, what can you expect from a station that gives us stock advice from Jim Cramer and bond advice from Rick Santelli? Betting advice from Steve Stevens fits right in to the programming mix. Where’s Stacy Keach when you need him?

Hi,

In Tennis it is easy to hit more than 67% bets correctly. Besides, in NFL I read this blog and it claims to have hit more than 70% accurate bets in the 2014 regular season of NFL. Here is the link – https://thefoolsbookie.com/blog/yes-humans-are-still-better-than-million-dollar-machines/ .

As a statistician I think the accuracy of number of bets does not make any sense in sports betting. Profitable sports betting depends on the odds that the bookmakers offer. If they offer odds like 1.1 then I will need a hit rate of more than 90% to make profits in the long run.

I would like to hear your feedback on this thought of mine, and the blog of the person.

Thanks

It looks like you’re using a standard deviation of 1.6%. If I’m reading that correctly, is that an observed STD?

Thanks,

Jason