Learn
The Two Side Test: how to find positive EV bets on both sides of a line
Here is how to find positive EV bets with the Two Side Test: get a fair price from another book, then price both sides of the line you're looking at against it. One line can be a good bet on one side and a bad one on the other. This worksheet walks through the video's hypothetical DraftKings example with the margin math filled in, then gives you a price ladder and the steps to run it on any line.
This is the worksheet from the video, in the same order as the video.
The test in three moves:
- Get a fair price first. A price only means something next to a fair price, so you take one from another book's two-sided market and remove its margin.
- Price both sides. In the hypothetical, the other book has -130 and +110, so the fair favorite is about 54 percent. At DraftKings, -110 on that favorite is worth about $3.61 per $100, and +110 on the underdog loses about $3.97.
- Remember the good side still loses. The favorite is a good price and still loses about 46 percent of the time.
All odds on this page are demo prices for illustration, not any sportsbook's actual line.
Step 1, get a fair price from another book
Take a two-sided market from a different book (ideally a sharp one with low margins) and turn each price into an implied chance. For a negative price that is the price divided by the price plus 100; for a positive price it is 100 divided by the price plus 100, the standard moneyline conversion.
| Other book (demo odds) | Price | Implied chance |
|---|---|---|
| Favorite | -130 | 56.52% |
| Underdog | +110 | 47.62% |
| Total | 104.14% |
The total is above 100 because of the margin. Divide each side by the total (the multiplicative method) and you get the fair chances: favorite 54.27 percent, underdog 45.73 percent. Other methods give slightly different answers; on a market this close to even they barely differ, but on heavy favorites the choice matters. Clarke, Kovalchik and Ingram found the power method universally outperformed the multiplicative method in their 2017 comparison.
Step 2, price both sides of the line
Expected value per $100 staked is the fair chance of winning times the profit, minus the fair chance of losing times the stake.
EV = fair win chance x profit - fair lose chance x 100
At the demo DraftKings line:
| Side (demo odds) | Price | Profit on $100 | Fair win chance | EV per $100 |
|---|---|---|---|---|
| Favorite | -110 | $90.91 | 54.27% | +$3.61 |
| Underdog | +110 | $110.00 | 45.73% | -$3.97 |
Same line, same game, one side positive and the other negative. The plus-money side looks like the bargain, but it is the one losing money here, because its fair price is about +119 and +110 pays less than that.
The price ladder
The fair chances give you a break-even price for each side: about -119 for the favorite and about +119 for the underdog. Anything better than that is positive EV, anything worse is negative. Here is the ladder for this hypothetical market, so you can see how fast the answer changes:
| Favorite price | EV per $100 | Underdog price | EV per $100 |
|---|---|---|---|
| +100 | +$8.55 | +125 | +$2.88 |
| -105 | +$5.96 | +120 | +$0.60 |
| -110 | +$3.61 | +115 | -$1.69 |
| -115 | +$1.47 | +110 | -$3.98 |
| -120 | -$0.50 | +105 | -$6.26 |
The tables use rounding, so a cent here or there can differ from a calculator.
The worksheet
Copy these rows for any line:
- Reference market: both prices from the other book.
- Implied chances: convert each price with the moneyline rules.
- Total: add them. Anything above 100 percent is margin.
- Fair chances: divide each implied chance by the total (or use power on lopsided markets).
- Target line: both prices at the book you want to bet.
- Profit on $100: 100 x 100 / price for negative prices, the price itself for positive ones.
- EV per $100: fair win chance x profit - fair lose chance x 100, for each side.
- Verdict: bet only a side with positive EV, and only if the edge is big enough to matter after rounding and line moves.
What to watch out for
- Your reference can be wrong. A fair price is only as good as the market it came from. Bookmakers tend to set a price and adjust it infrequently, so a stale reference line can make a gap look real when it isn't.
- Positive EV still loses. In this example the good side loses about 46 percent of the time. Over 100 bets of $100 at +$3.61, the expected profit is about $361 but the standard deviation is about $951, so a losing stretch is normal and expected.
- Winning gets noticed. Researchers who used public odds to find mispriced bets and staked real money for five months reported that the industry responds to successful clients with discriminatory practices such as limits. In Massachusetts, regulators found 0.64 percent of accounts limited over a recent period and voted on rules to tell bettors when and why.
How to tell it worked
Run the hypothetical through your sheet: if you get 54.27 percent, +$3.61 and -$3.97, it is set up right. On real lines, track each bet's price against the closing price; closing line value is positive when your price beat the close. A process that finds real positive EV should get better prices than the close more often than not, long before the win-loss record means much.
Try the numbers in the expected value calculator and the devig calculator, and see the DraftKings positive EV bets page for how this applies to one book. The Quant+ board shows each line's best price, the de-vigged fair value and the edge side by side.
Quant+ is analytics for informational purposes. 21+. Nothing is guaranteed. If you or someone you know has a gambling problem, call 1-800-GAMBLER.
Questions people ask
What is a positive EV bet?
A bet whose price pays more than its fair chance of winning is worth, so its expected value per bet is above zero. It can still lose any single time; positive EV only describes the long-run average.
Where do I get a fair price?
From another book's two-sided market with its margin removed, ideally a book with low margins whose prices move quickly. Some bettors average several books.
Can both sides of a line be positive EV?
Only if the book's line is far off the fair price on both sides at once, which is rare because the book's own margin usually makes at least one side negative. Usually at most one side is a bet.
How big an edge is worth betting?
That depends on how confident you are in the fair price and how much the line can move. Many bettors skip edges of a dollar or two per hundred because small errors in the reference price can erase them.
Sources
- Moneyline odds, Wikipedia
- Adjusting bookmaker's odds to allow for overround, American Journal of Sports Science
- Why are gambling markets organised so differently from financial markets?, Economic Journal via RePEc
- Beating the bookies with their own numbers, arXiv (Kaunitz, Zhong, Kreiner)
- Expected value, Wikipedia
- Closing Line Value (CLV), SharpAPI docs
- Massachusetts Bettors Could Soon Have Wagering Limitations Explained, Covers
