How We Bet

A system is just a historical situation that has produced an edge over a meaningful sample. It is not a crystal ball and it is not a guarantee. It is a starting point. When a system fires on a game, it tells us that teams in this exact spot have historically performed a certain way against the number, and that is enough to make us stop and look closer.

Our main recommended source for historical trends is Action Network Bet Labs. It lets you build and test systems across years of data with the kind of filters that actually matter. If Bet Labs is out of your budget, Trend Labs is a cheaper alternative that still lets you research historical angles. It is not our top choice, but it is a real option for someone getting started.

We are also building our own historical trend tool, built on the same data and signals you see across the product, and mentorship members get access to it as part of their membership. The idea is simple: instead of paying for a separate database subscription, you research your angles inside the Insider ecosystem and carry them straight onto the live board, with our team there to help you build and pressure-test the systems that fit your bankroll.

Here is the single most important question to ask about any system: does it make sense? You can throw enough random filters at a database and produce a 65% record out of pure noise. That number looks great and means nothing. The trends that hold up are the ones that have a logical reason behind them, a narrative you can actually explain.

Revenge spots make sense. A team coming off a blowout loss covering the next game makes sense. A desperate team in an elimination spot making sense. These angles describe real human and competitive behavior, not coincidence. If you cannot explain why a system would work without pointing at the win rate, be very skeptical of it.

When you are evaluating a trend, separate the result from the reasoning. A strong-looking record built on a logical, repeatable situation is worth far more than a stronger-looking record built on filters that have no business being connected.

Sample size is the biggest lever in trend research, but raw sample size without context is misleading. A 25-game sample spread across five random regular seasons is usually garbage. A 25-game sample concentrated in the playoffs can be genuinely valuable, because playoff conditions are more consistent and the spots carry more weight. The size matters, but so does what the sample actually represents.

Do not sleep on ROI. Record and hit rate get all the attention, but ROI tells you whether a system is actually making money relative to the prices you are paying. A system can hit at a high rate and still be unprofitable if it only wins on heavy favorites, and a system can hit at a lower rate and crush if it wins on plus-money dogs. In a lot of cases, ROI is just as important as the record.

And when a historically strong system goes cold, reduce your confidence in it. Hot and cold streaks are real. There is no reason to force a trend that is clearly out of form just because it looked great over a long sample.

The other major place a bet starts is the market itself. Line movement is information. Where a number opened, where it sits now, which direction it moved, and how it moved relative to the betting action all tell a story about who is on what side and how seriously the book is taking it.

We pay attention to movement against the public, movement that confirms the public, steam moves, and the timing of moves. There is no single type of movement that matters most. It is all situational. The same move can mean very different things depending on the game, the market, and what the public is doing.

Line movement and systems work together. Sometimes strong line movement pushes us off a bet that had great systems behind it. It is actually easier for a strong system to push us off a pure line-movement read than the other way around, but neither is an absolute. When systems and line movement do not align at all, we tend to stay away. There is no need to force anything.

Inside the product, we surface three signals that summarize what the market is doing. Pressure Move means the majority of bets and money are on a side and the line movement is confirming that public lean. Vegas Backed means a side has the minority of bets and money, yet the line is still moving in its favor, which is the classic fingerprint of sharp, respected action.

Whale Respect means the money percentage significantly exceeds the bet count, with supporting movement, which points to large individual wagers driving the number rather than the crowd. Each signal tells you something different about where the meaningful money is, and each is useful in different situations.

None of these signals is an automatic bet. They are inputs. A signal is most powerful when it lines up with a system, a logical narrative, and a number that still has value. On its own, a signal is just one piece of the picture.

Public betting data shows you two different things: the percentage of bets (tickets) on a side and the percentage of money on a side. When those two numbers diverge, it is telling you something. A side with a small share of tickets but a large share of money usually means a smaller number of larger, more confident wagers, which is worth respecting.

The lazy version of this is to blindly fade the public, and that is not what we do. The public is right plenty of the time. The point is to understand the market, not to reflexively bet against the crowd.

The most useful tell is when there is heavy public support on one side and the line is not moving the way you would expect. That disconnect is often a sign that you may be looking at a trap, or at least a number the book is comfortable taking action on. That is your cue to slow down and dig deeper.

All +EV means is 'expected value' — a bet priced better than its true probability says it should be. People find +EV bets in a few different ways. Some build their own projection models, which we find clunky and, frankly, a little stupid: every sportsbook already runs models with more data and more resources than you will ever have. The other way is to base it off the sharp books, and that is exactly how we do it.

Here is the honest part: most of the bets you see us take are already +EV. We just feel no need to mention it every time, because for us it is the standard, not the selling point. Even the slightest edge in expected value matters over a long sample — but EV alone is not what wins the bet. It is the price filter every play has to pass through before the rest of the read even matters.

How we calculate it is simple and transparent. We stream Pinnacle and Circa odds in real time, average the two, or use whichever one is available if a market only exists at one of them. After multiple backtests against where games actually ended up and where the closing line settled, those two books are by far the sharpest on the board. So if you are getting a meaningfully better price on your bet than what Pinnacle and Circa are offering, that is +EV. It is far from a guarantee — but it is one of the best starting points in all of betting.

Narrative is not fluff, and it is not fan talk. It is a real situational layer. The word says it: anything that is genuinely a sports narrative can play a role. Revenge spots, a key player facing his former team, a coach against the team that fired him, a rivalry game, an embarrassing loss that demands a response. These show up constantly, and they matter.

Narrative carries the most weight in the NFL and in the playoffs, because those games hold more meaning and the human element is amplified. A revenge spot in Week 12 of the NFL or a desperate team down in a playoff series is exactly the kind of situation where motivation and emotion actually move the needle.

The key is that narrative has to connect logically with the rest of the read. We are not betting a revenge spot just because it sounds good on television. We are looking for a narrative that lines up with a system, the market, and the matchup. When all of those agree, the narrative is no longer a story, it is a reason.

This is where everything comes together. A bet starts from a system or from the market. From there we ask whether the system makes sense, whether line movement and signals support the same side, whether the public is falling into something obvious, whether there is a real narrative, and whether the number is still playable.

The more of those that point in the same direction, the more interested we get. Full alignment is rare, and you do not need it. But you do need enough agreement that the bet is not fighting against its own signals. When systems and the market flatly disagree, that is usually a pass.

Alignment is also what determines unit size and whether we take the bet at all. A play with strong, broad alignment might be a one-unit bet. A thinner read might be a half unit or a pass. We will get into unit sizing in detail later, but the principle starts here: the bet is only as strong as the agreement behind it.