Original research · August 11, 2026
A One Point Spread That Actually Costs 4.25 Percent
By Matt Downs
There is a check that takes about ten seconds and almost nobody runs it. Take any two-sided market, write down what it costs to buy each side, and add the two numbers. Anything over 100 is what the house is charging you to be right.
Run it on a prediction market and you get a number that looks unbeatable. Run it again with the fee included and you get a different number. This is the arithmetic on a real market, pulled on the morning of August 11, 2026.
How to calculate the hold on any market
A two-sided market gives you two prices. Convert each one to the probability it implies, add them, and subtract 100. What is left over is the overround, the margin the venue has built into the pair.
A standard sportsbook price of -110 on both sides implies 52.38% twice. That adds to 104.76%, so the book is charging 4.76 points. This is the number people are talking about when they complain about juice, and it is the baseline every other venue should be measured against.
You do not need to do the conversion by hand. The free no-vig calculator takes both American prices and prints the vig along with the fair line underneath it.
What Kalshi's board quoted on Dallas at Seattle
On the morning of August 11, 2026, Kalshi's market on the Dallas at Seattle NFL preseason game was quoting Seattle at 64 cents and Dallas at 37 cents. Those add to 101, so the quoted cut is a single point.
In American odds that is -178 and +170. The calculator reads the pair at implied 64.03% and 37.04%, for a vig of 1.07%. Against a sportsbook's 4.76%, that looks like it settles the argument.
The Kalshi fee formula, and why it is not in the quote
Kalshi charges a trading fee on top of the price, and it does not appear in the number on the screen. The published taker formula is 0.07 x C x P x (1-P), where C is the contract count and P is the price in dollars, rounded up to the nearest cent once per order.
On 100 contracts of Seattle at 64 cents that is 0.07 x 100 x 0.64 x 0.36, which is $1.62. On the Dallas side at 37 cents it is $1.64.
So you are not paying 64 cents, you are paying 65.62. The other side is not 37, it is 38.64. Add the two real numbers and you get 104.25, not 101. Entered as -191 and +159, the same calculator reads the vig at 4.25%.
The number to sit with: a market advertising a one point spread charges 4.25 points, and the sportsbook everybody complains about charges 4.76. Half a point apart. The fee is not a rounding detail on a prediction market, it is most of the cost.
The same bet, priced three ways
| Kalshi, quoted price only (64c + 37c) | 1.07% vig | |
|---|---|---|
| Kalshi, taker fee included (65.62c + 38.64c) | 4.25% vig | |
| Sportsbook at -110 both sides | 4.76% vig |
The cut is set by the market, not by the logo
Three markets pulled off the same board within the same hour told three different stories. The NFL markets quoted one point. The MLB games quoted two and three. The Los Angeles at New York WNBA game quoted eight.
Same venue, same fee schedule, same morning. What changed is how many people were trading each game. A thin market has a wide spread because nobody is standing there competing to fill you, and the width comes out of your side of the trade.
This is why a venue-level verdict is close to useless. Check the market you are about to trade, not the reputation of the app it sits in.
Why longshots pay a much higher fee rate
The P x (1-P) term peaks at 50 cents, so the fee is largest in raw dollars on a coin flip. As a share of what you actually put up, though, it gets steadily worse the cheaper the contract is.
At 20 cents, the fee on 100 contracts is $1.13 against $20 at risk, which is 5.65%. At 88 cents it is $0.74 against $88, which is 0.84%. A longshot player is paying close to seven times the rate a favorite player pays, and nothing on the screen mentions it.
One qualifier that matters: these are taker fees, what you pay for hitting a price already resting on the book. Posting your own limit order is charged at a lower rate, so everything above is the worst case.
Frequently asked questions
How much does Kalshi charge in fees?
The published taker formula is 0.07 x C x P x (1-P), where C is your contract count and P is the price in dollars, rounded up to the nearest cent. On 100 contracts at 64 cents that is $1.62, and the maximum is $1.75 per 100 contracts, which occurs at 50 cents. Maker orders that rest on the book are charged at a lower rate.
Is Kalshi cheaper than a sportsbook?
On the quoted spread, usually yes. All in, often not. A market quoting 64 cents and 37 cents advertises a 1.07% vig, but adding the taker fee to both sides produces 4.25%, against 4.76% for a standard -110 sportsbook price. The answer depends on the specific market and how heavily it is traded, so run the numbers on the market you are about to trade.
What is a good hold percentage?
Lower is better and the useful reference points are these: 4.76% is a standard -110 two-way sportsbook price, anything at or under about 2% is a genuinely sharp two-way market, and anything above about 6% means you are paying a lot for the convenience. Thin markets in low-volume sports routinely run 8% and higher.
How do I calculate hold myself?
Convert both sides to implied probability, add them, and subtract 100. The remainder is the overround. Include any per-trade fee in the price before you convert, or you will measure the advertised cost instead of the real one.