Skip to content
Notes

Sharp Players: How to Spot Them and What to Do

What a sharp player looks like on your sheet, why a winning streak is not the same thing, and the moves that protect the margin without burning the account.

By
Mike Casanova
Published
Checked
In this note
7 min read

A sharp is not a player who is winning

Every book has somebody who is up. Most of them are not sharp. They are a few weeks into a run that will turn, and an operator who reacts to the run instead of the method ends up cutting his best customer and keeping the one who will quietly take him apart.

A sharp player is one whose bets carry information your number does not have yet. The result is the last place that shows up. The timing, the size and the choice of number show up first.

What the research says about streaks

This is worth getting straight before you start reading tells into a sheet, because the instinct to see skill in a winning month is strong and usually wrong.

Steven Levitt studied one season of wagers placed at an online sportsbook, 19,770 bets from 285 bettors picking five games a week for seventeen weeks. The spread of results across those bettors matched what he described as independent tosses of a 50-50 coin. How a bettor did in the first half of the season predicted nothing about the second half, and the predicted win rates clustered between 49 and 51 percent.

So the base rate is this: the overwhelming majority of players who are up right now are not sharp. They are the tail of a distribution. That is not a reason to ignore the ones who are. It is a reason to ask what the player is doing rather than what he has made.

The same study found something more useful for an operator. In almost half of all games, at least two thirds of the money landed on one side, and in roughly one game in ten more than 80 percent of it did. Books in that data were not balancing anything. They were pricing to where the public wanted to go and keeping the difference, which is also why a lopsided sheet is normal rather than an emergency.

How a sharp actually beats a book

The popular picture of a sharp is a man with a model who knows which side wins. The documented version is duller and more dangerous.

In 2017 Lisandro Kaunitz, Shenjun Zhong and Javier Kreiner published an experiment in which they never tried to predict a result. They took the average price across many books as a stand in for the true probability, then bet only where one book's number sat far enough outside that consensus. Over five months of real money at a fixed stake they placed 265 bets and returned 8.5 percent, winning 47.2 percent of them.

Read that again, because it is the whole lesson. They won fewer than half their bets and still made money, because they only ever bet when the price was wrong. The attack surface was not the game. It was the stale number.

Your book has exactly the same surface. A number you have not moved since Thursday, a second half line that lags the live feed, a prop you copied once and forgot: that is what a sharp is shopping for. He does not need to be smarter than you about the game. He needs you to be slower than the market.

The signs on your own sheet

None of these is proof on its own. Two or three of them together on the same player, over a month, is a pattern.

He bets within minutes of your number going up, or within minutes of news. Early action is aimed at a price you have not finished thinking about. Late action is aimed at one you have not updated.

Your line moves the way he bet, by itself. If the market agrees with him an hour after he bets, he is not guessing.

His stake changes with the price and not with his mood. A recreational player bets about the same amount on everything he likes. A sharp bets a little most weeks and a lot on the one number he thinks is broken.

He takes the half point rather than the team. A player who waits for a key number, or who only bets the side after it crosses a particular figure, is reading the price as a price.

He has no favorites. Nothing in his history says he lives in your town or likes anybody. He bets whichever side is mispriced, which is the single hardest thing for a recreational player to fake.

What the big books do with players like this

They limit them, and they are not shy about it. The 2017 experiment ended because the books caught up: the authors reported that within months their stakes were capped, in several cases to a fraction of their original size, some wagers were routed to manual inspection before being accepted, and others were refused outright. The strategy still worked. It just could no longer be run at any size worth the trouble.

That is the whole toolkit, and a small book has every piece of it. The difference is that a large operator can cut a sharp to pocket change and feel nothing, while you probably want the account. A player capped to a couple of dollars does not hang around.

What to do, in order

Start by sizing, not by closing. Set the limit on that player, for that sport, at a number where a correct bet costs you a normal week rather than a bad month. Setting a player's weekly credit works through how to pick the figure, and the per player and per sport controls are where it gets set.

Then slow him down. A short approval step on anything above a threshold costs a recreational player nothing, because he is betting the game he was going to watch anyway. It costs a sharp his whole advantage, because his advantage has a shelf life measured in minutes.

Then move your number when he bets. His action is information whether you want it or not, so use it. Reading line movement covers how to tell a move driven by money from a move driven by news.

Only then consider the account itself, and do it on the terms you published rather than after a result you did not like. Voiding a settled bet because it turned out badly is the one move that loses players you wanted to keep, and it travels: every other player in town hears about it by Tuesday.

Write the rules down before you need them

A licensed book does not get to improvise this. Nevada requires every sports pool to adopt, post and follow written house rules covering what is paid on winning wagers, what happens when a schedule changes, how long a winning ticket can be redeemed and how patrons learn about a line change, and to state where wagers may be accepted at terms other than those posted. Under the same regulations the state can require a book to disclose its betting limits and get approval before changing them.

Nobody is going to make you do that. The questions arrive anyway, usually from the player you least want to argue with. An operator with the answers written down has a short conversation. An operator deciding case by case has a long one, and the player always remembers the call that went against him.

One week with one sharp

Call him Rubén, in a town called Palmares, on a roster of forty.

Everyone else bets between 50 and 200 a game and takes whatever number is up. Rubén bets 100 most weeks. Three times this season he has bet his full limit inside ten minutes of a number going up, and each time the market moved his way within the hour.

On a straight price of -110 the margin on a balanced pair of bets is 4.55 percent of what is risked, so forty players at a few hundred a week is a figure you can plan around. One player allowed to bet ten times the house average, hitting only the numbers that are late, is not in that plan. He is not beating your judgment of the games. He is beating your update speed, and the fix is the same either way: cap him, slow him, and move the number when he moves.

Cut his limit to the roster average and he is a line feed you do not pay for. Close him and you learn nothing, and the next one takes longer to find.

What the platform does and does not do

BookiePerHead is the platform and the service behind a book. It takes no wagers and holds no player money, so limits, approvals and pricing are the operator's calls, made in the agent tools. What the software does is make those calls fast enough to matter, and keep a record of who bet what and when, which is the only evidence that ever settles the question of whether a player is sharp.

The bill follows active players rather than how much any one of them bets, so tightening a limit is a risk decision and never a billing one.

Rules on who you can take action from, what you may collect and how any of this is taxed depend on where you and your players are. That is a question for a lawyer in your jurisdiction, not for a blog post.

01Questions

Asked about this

Look at what he bets, not what he wins. A sharp takes the same side you were about to move, bets early off a stale number or late off news, and sizes up only when the price is wrong. A lucky player bets the same amount on everything and takes whatever number is showing.

Rarely. A closed account tells you nothing and sends the player to another book. A smaller limit, a short approval delay and a price that moves as soon as he bets keeps the account and takes away most of the edge.

Weeks, not days. One season of wagering data at an online sportsbook produced a spread of bettor results that matched coin tosses, and past results predicted nothing about future ones. Short runs are noise in both directions.

Yes, mostly through size rather than frequency. One player allowed to bet several times what everyone else can turns a single game into the whole week, which is the argument for a limit per player and per sport.

Limits, what happens on a postponed event, how long a winning ticket is good for, and that a wager can be taken at terms other than the ones posted. Regulated books are required to publish rules like these. A private book has the same arguments without the regulator.

Often. A sharp who bets a capped amount is the cheapest line feed you will ever have, because his action tells you which of your numbers is behind the market before the rest of the roster finds it.

02Next

More notes

  1. 7 min read

    Teasers, Pleasers and Buying Points, Explained

    How a teaser moves the line, what those points really cost, when a half point is a leak for the book, and the teaser rules to write down before the season.

  2. 6 min read

    Cash Out, Explained for Operators

    What cash out is, how the offered price is built, what it does to an open position, and the house rules to publish before a book turns it on.

  3. 6 min read

    How Parlays Are Priced, and Why Books Offer Them

    A parlay multiplies the legs, so it multiplies the margin. The arithmetic, where a fixed card disagrees with it, and the settings an operator should check.

All notes

03Get a demo

Read enough? See the platform.

Tell us about your book. We reply with a demo login and a time to walk through it on real screens.

On the call

  • A demo login for the player site and the agent backend.
  • Whatever this note left you asking.
  • Your weekly price, and how you pay.

We reply within 24 hours.

Outside American football season and the major playoffs there is a free trial as well; we go over it when we talk.

Are you running a book already?
Preferred language

No long-term commitment. We reply to every request ourselves.