How To Spot A Fraud
The Guy on Your Timeline
You know the guy. He is on TikTok, Twitter, Instagram, maybe all three. Intense energy. All-or-nothing takes. “Lock of the century.” “We do not lose.” Screenshot of a four-leg parlay that hit for $8,000. Link in bio. First week free.
He might be 22. He might be 55. The personality is always the same: supreme confidence, flashy results, and zero explanation of how any of it works.
Here is the first question you should ask, and almost nobody does:
Not “do they seem confident.” Not “did they go 4-0 last Saturday.” What is their actual, demonstrable, trained skill?
Because the NFL betting market is not some casual game. It is priced by sportsbooks that employ teams of quantitative analysts with graduate degrees in statistics and mathematics. It is moved by billions of dollars from professional syndicates who have been doing this for decades.
The number you see on your screen is the output of an enormous amount of math, data, and sharp money.
Beating that market requires the same type of skills the books use to set the lines: statistical modeling, predictive analytics, probability theory, and working with noisy data under uncertainty.
These are skills that come from formal training and professional experience, not from watching a lot of football.
The Income Test
Here is something that should immediately make you skeptical of most pick sellers.
The quantitative skills required to build a winning NFL model are the same skills that pay $200,000 to $400,000 a year in industries like pharmaceuticals, finance, and tech.
Companies pay that much because people who can build statistical models, run simulations, and make evidence-based decisions under uncertainty are rare and valuable.
Now do the math on what it takes to replace that income with sports betting.
A strong NFL model wins at roughly 55-57% against the spread. That translates to about a 5-7% ROI on every dollar wagered.
To generate $300,000 in income from a 7% ROI, you would need to wager more than $4 million per season.
At 230 bets across a season, that works out to roughly $18,000 per game.
You cannot simply bet $18,000 on every NFL opening line. The liquidity is not there on many openers, and once you start betting significant size, sportsbooks can limit or ban your account.
So here is the reality: someone with the skills to actually beat the NFL market can often make far more money using those skills in a professional career than they can betting.
The betting is a supplement, not a replacement. It is a second income stream built on the same toolkit they use in their day job.
Now look at the guy on TikTok. If he had the quantitative skills to beat the NFL, he would likely have valuable professional applications for those same skills. He would not necessarily need to rely on selling $25 pick packages for his livelihood.
That is not a foolproof test. But when someone's entire visible income appears to come from selling picks, and they show no evidence of the training or career that would produce the skills to generate those picks, you should be asking hard questions.
The person behind BTB's models has doctoral and graduate-level training in health economics and quantitative methods, with a full-time career at the highest level of the pharmaceutical industry building statistical models, running simulations, and synthesizing real-world evidence.
The NFL model is a direct application of the same toolkit. That is not “I watch a lot of football and I know ball.” It is formal training in the exact discipline that drives modern sports betting.
What an “Edge” Actually Is
Most people hear “edge” and think it means you know who is going to win the game. That is not what it means. Not even close.
Here is a way to think about it that does not require a statistics degree. The best NFL prediction models in the world, including the ones the books use, are wrong by about 13 to 14 points per game on average.
That is not a flaw. That is the floor.
When you have 22 players on the field, along with injuries, weather, play-calling, turnovers, and countless other variables, there is a minimum amount of error that cannot be eliminated.
No model, no matter how good, is going to predict the final margin of an NFL game within three or four points consistently. The real world is too chaotic.
So what does a winning model do?
It gets that average error slightly lower. Instead of being wrong by 13.7 points per game, it might be wrong by 13.3.
That sounds like nothing. But across hundreds of bets, that small improvement in accuracy can be the difference between losing money and generating a meaningful return.
BTB estimates its true edge at roughly 3%.
On a standard -110 spread, the breakeven win rate is approximately 52.4%. We are generally targeting an actual probability around 55-56%.
Three percent. That is it.
That is what a real edge looks like in an efficient market.
If someone tells you they have a 65% long-term NFL win rate or that they “never lose,” you should be extremely skeptical.
The market is too efficient for edges that large to persist over meaningful sample sizes. A 3% edge sustained over multiple seasons can be genuinely elite. And it still produces plenty of losing bets.
The Metric That Matters Most: Closing Line Value
This is one of the most important concepts in sports betting, and many recreational bettors have never heard of it.
The closing line is the final number available before kickoff. After a full week of betting action, sharp money, injury news, and model updates, the closing line represents the market's best estimate of the true price.
It is generally the most informed version of the market for that game.
Closing Line Value, or CLV, measures whether you got a better price than the eventual closing line.
Suppose you take the Chiefs at -3 on Tuesday.
By kickoff, the market closes at Chiefs -4.5.
You captured 1.5 points of CLV.
The market moved toward your number. You were able to get a better price before the rest of the market adjusted.
CLV is a powerful signal that an edge may exist.
Here is why: win rate over 20, 50, or even 100 bets can still be heavily influenced by luck. You can flip a fair coin and get an unusually high percentage of heads over a relatively small sample.
But if you consistently get better numbers than the closing market over hundreds of bets, that provides evidence that your process is identifying value before the market fully corrects.
And this part is critical:
You will make bets with strong positive CLV and lose them. You will make bets where the market moves against you and still win.
On any single game, CLV tells you very little about the final outcome. Over 200 games, however, consistently positive CLV tells you a great deal about whether your process is sound.
Think of it this way. Imagine you know a coin is weighted to land heads 53% of the time. Someone offers you +105 odds on heads. That is a favorable price. You take it. The coin lands tails.
Did you make a bad decision?
No. You made a mathematically favorable decision that happened to lose.
You would take the same opportunity again because the price was favorable.
CLV helps answer the same question in betting: did you consistently get the right price?
When you are in a losing streak and everything feels broken, CLV can help determine whether the model is actually deteriorating or whether you are experiencing normal short-term variance.
How to Verify Anyone, Including Us
We are not asking you to trust us. We are asking you to verify.
Here is exactly how.
- Use a tracking platform. BetStamp and Pikkit allow bettors to timestamp entries, compare entry prices to closing lines, and evaluate long-term records.
- Ask for CLV, not just win rate. A short-term winning percentage can be heavily influenced by variance. A large sample with persistent positive CLV is much more informative.
- Ask for a full-season record. Not a highlight reel. Not the best week of the season. Every win, every loss, and every wager.
- Run from guaranteed outcomes. “Lock.” “Guaranteed winner.” “Can't lose.” Nobody who genuinely understands probability and variance should guarantee the result of an individual game.
Screenshots of winning tickets are not verification. Tracked, timestamped results are.
A 60% record over 30 bets tells you very little. A strong record over 200+ bets combined with consistent positive CLV tells you far more about whether a process has a real edge.
People who understand betting markets tend to talk about process, sample size, expected value, probabilities, and closing line value.
If the language sounds more like a lottery commercial, you should probably treat it like one.
Required Reading
If anything in this piece made you think differently about how sports betting works, there is one book that will take you much further: The Logic of Sports Betting by Ed Miller and Matthew Davidow.
It goes deeper into market efficiency, how sportsbooks set lines, closing line value, why most bettors lose, and how the small percentage who win actually approach the market.
It is not a system book. It will not give you picks.
It gives you a framework for evaluating every service, every tout, and every claim you encounter in this industry.
What This Looks Like in Practice
Everything above is theory.
What does a real, verified, CLV-positive edge look like over three NFL seasons? What does +22.6 units and a 7.1% ROI actually feel like week to week?
It feels a lot worse than most people think.
And that is exactly why understanding variance, process, and market pricing matters.