There is no such thing as being bad at betting against a -110 line
Betting against a -110 line is, for a majority of people, an exercise in mediocrity. It is uncommon to be anything other than a 4.76% punter.
I assume everyone reading this understands that to be profitable against a -110, you must win more than 52.38% of your wagers. That is significantly harder than it seems in absolute numbers. The good news is that it is equally as hard to be bad. To be truly bad, which ironically makes you good, you must win less than 47.62% of your wagers. If we subtract the difference, we are left with a range of 4.76%, in which 90% of the population resides.
If someone managed to fall outside the range in either direction, they are quite good. In the first instance, if you win with a greater frequency than 52.38%, simply bet your sides. If you win less than 47.63% of your wagers, simply fade yourself to profitability.
There is no such thing as being bad at betting on sports. There is mediocre, in which case you lose money. And very good, in which case you make money.
How Good Can Someone Be Against A -110 line?
Not great in absolute terms. You will see people boast all sorts of win percentages. Ahem. The problem, assuming their veracity, is the sample size.
Suppose you bought a box of raisin bran cereal. And you poured out four bowls of cereal in precisely equal measure. Each bowl is most likely to have a different amount of raisins in it, despite the box from which they all came having a constant ratio of flakes to raisins. Wins in sports betting operate identically. While probabilities are fair in their potentialities, they are not fair in their outcomes.
When, for example, Kevin Durant takes an off-the-dribble three from the right-wing, there is an exact probability that he will make the shot. However, that particular shot will either be successful or not. It will resolve to a single outcome. It is in a state akin to superposition until he makes or misses it. In binary trials, there are winners and losers. Moreover, wins and losses do not follow in an orderly sequence. Like the raisins in your cereal box, they come in streaks, bunches, and clusters.
For example, I created a Twitter account last January to document the model’s prediction on the Super Bowl. I do not use the account anymore, but I was able to look up the tweet. The model came out with an exceptionally high degree of confidence in the game. It also came out with five sides, Tampa Bay +3.5, Under the total, both quarterbacks under their passing totals, and Brady for MVP.
Of course, five straight wins does not make the model 100% against a -110 line. (I realize the MVP would have been longer than -110, but since it is longer, the point perseveres even without the exact number) If I add those plays to the first week of the season this year, the model was 12-1 in its last thirteen games. That does not make the model 92.3% against a -110 line either. Finally, if I add those five wins to the first two weeks of this season, the model has gone 14-4 in its last 18 games. Yet, it does not follow that the model is 77.7% against a -110 line. The sample size is too small to reach any conclusions.
This season’s Westgate Super Contest winner is likely to have a win percentage in the mid to low 70% range. (If you do not know, the contest requires entrants to pick five games per week in the NFL against a -110 line) It does not mean that the winner is 70(x)% against a -110 line. It means that their bowl got the most raisins this season. It is possible to perform that well in a sample size of 85 games. It is not possible to perform that well over a sample size of 850 games.
Rather than show you the math to prove this, I will say the illustrious contest has only ever had one person claim first place multiple times. And they have won only twice. Given that the entry fee is $1,000, and first place takes home a million, this contest should be a yearly ATM that pays 1,000 to 1 to anyone consistently correct 70(x)% of the time. It isn’t because those people do not exist.
Given a sufficient sample size, we can deduce that the most prolific bettors against a -110 line perform under 70%. We also know that a successful bettor must win at least more than 52.38% of such wagers. That still leaves us a discrepancy of 17.62%.
If we turn to the Westgate Super Contest again, we can refine the number. Looking at the number of entrants each year and the percentage that finishes the contest in the low 70 percentile, we can use a binomial distribution to determine the most likely actual win percentage for the best performing individuals.
The most probable win percentage for those individuals over a sufficient sample size is 62%, ranging between 61% and 64%. The best sports bettors in the world are successful on 62-64 percent of their wagers.
A volume of 100 bets in an NFL season against a -110 line results in an 18.3% return on investment over the season. Thus a person with that percentage who bets 100 games in a season at $1,000 units (Vig Excluded) will risk $110,000. They will return $130,200. They will experience a net profit of $20,200, which is 18.3%. (Some prefer to base ROI on bankroll size. A person betting $1,000 units should have a bankroll between 35-40K. They would claim an ROI of roughly 50%)
It is not a bad ROI for six months. However, the numbers are not that simple. As your bankroll grows, so too should your unit size. You will also be utilizing a Kelly approach that will maximize your profits further. Those are baseline numbers only.