LongVol/ShortVol Fee Schedule
LongVol and ShortVol should pay a small fee every time someone trades them, rather than the yearly fee on balances that RiskON and RiskOFF pay. These tokens turn the whole deposit into a new bet every week, so the fee belongs on that weekly bet. I recommend that The Risk Protocol (TRP) use the fee formula of Kalshi and Polymarket: long-volatility trades pay their 7% taker rate, and short-volatility trades pay a quarter of that, 1.75%, Kalshi's maker rate, because ShortVol puts up the money, is there for the yield and is more sensitive to fees.
Buying LongVol, or keeping LongVol by selling ShortVol
Charged on each token traded, it comes to about 6% of the LongVol ticket.
With BTC at $100,000, a LongVol token costs $1,647.83 and its buyer pays $102.68 on top.
Buying ShortVol, or keeping ShortVol by selling LongVol
Charged on each token traded, it comes to about 1.5% of the premium, the money ShortVol gets for its LongVol each week.
Selling that LongVol token to keep the ShortVol costs $25.67.
p = $1,647.83 ÷ $15,000 = 0.109867% × 0.10986 × 0.89014 × $15,000 = $102.68
When the fee is charged
- Mint
A depositor puts in $15,000 and gets one LongVol and one ShortVol, which together are always worth the deposit, so holding both carries no risk.
Free - Sell one side
The depositor keeps the ShortVol and sells the LongVol for $1,647.83.
Depositor $25.67, buyer $102.68 - Hold
Balances carry no fee, and sending a token to another wallet is not a trade.
Free - Settle
On Friday the deposit is split between the two tokens by where BTC ends. The holders get back $15,000 between them, $128.35 less than they paid in, which is the fees.
Free - Rebase
Each holder's settled value comes back as new pairs for the next week.
Free - Roll
To stay on one side, a holder sells the other half of its new pairs, by hand or through the opt-in auto-roll.
Same as any sale - Redeem
Or the holder returns its pairs for USDC.
Free
A depositor who keeps the LongVol and sells the ShortVol pays the same $102.68 fee as a LongVol buyer, so its LongVol costs the same $1,750.51, even though the ShortVol it sells is worth about eight times as much. The fee depends only on p × (1 − p), which is the same for both tokens because ShortVol's share of the deposit is 1 − p.
On the trade screen
Traders see all-in prices, with the fee folded in and no separate fee line. A buyer pays $1,750.51 for one LongVol and a seller gets $1,622.16, on either side of the token's value before fees, its Net Token Value (NTV) of $1,647.83.
Prices include all fees and refresh every 15 seconds.
a year in fees per $1M deposited, in the BTC backtest for 2024–26 with each LongVol ticket resold once a week, or $436k if no ticket is resold.
is how each week's fee splits. ShortVol posts the collateral and is there for the yield, so it pays the smaller share.
of ShortVol's average profit is left after it pays its fees, in the same backtest.
of deposits stay when these fees are charged, the median of 400 Monte Carlo runs of the market, each run with and without the fees.
Why the RiskON/RiskOFF fees do not fit
RiskON and RiskOFF charge 2% a year on balances, taken daily, plus 15 basis points (bps) on each swap. Those fees suit tokens that people hold for months. LongVol and ShortVol differ in three ways.
The deposit is bet again every week
With $1M deposited and LongVol at 10% of the deposit, ShortVol holders sell $100,000 of LongVol every week, while a 2% yearly fee on the same $1M collects about $385 that week.
A cheap ticket to a big payout
With BTC at $100,000, a $1,647.83 token can pay the whole $15,000 deposit. A 15 bps swap fee takes $2.47 of it, while Deribit charges $60 for the same payout built from options1 and Kalshi or Polymarket $102.68 for a yes-or-no bet at the same price.2, 3
The money sits with ShortVol
ShortVol owned about 89% of each BTC pair and 78% of each ETH pair at the open in 2024–26, so a fee on balances falls mostly on the side whose money the market needs, and barely touches LongVol buyers, whose demand is inelastic.
Who will use it, and how often
Over the 350 weeks from January 2020 to September 2026, each opening at 08:00 UTC on a Friday, the two tokens behaved very differently. Every week is priced with our own risk engine's volatility forecast.
| ShortVol, the side that puts up the money | LongVol, the side that buys the payout | |
|---|---|---|
| Who | Option sellers and range liquidity providers after a weekly premium, desks hedging on Deribit, and investors using a slice of a portfolio | Traders betting on big moves or news, prediction-market bettors, and liquidity providers hedging their ranges |
| How it pays | Won 73% of BTC weeks and 69% of ETH weeks; lost everything in 15 BTC and 24 ETH weeks, about one week in 23 on BTC and one in 15 on ETH, most of them in 2020–21 | Worthless in 60% of BTC weeks and 47% of ETH weeks, and paid the whole deposit in 4% and 7%; at most 6.8 times its price in a typical BTC week and 3.5 times on ETH, before the fee |
| How long it is held | Weeks to months, rolled every week | One week at most |
| How often it trades | Once a week, at the roll | Bought weekly and traded during the week; in 35% of BTC weeks it doubled at some point, a median of 3.3 days after the open |
| Minting and redeeming | Mints pairs weekly to stay fully short volatility; redeems on the way out, often after a heavy loss | Winners redeem their pairs for USDC each Friday |
| Sensitivity to fees | High, as every fee comes straight out of its yield | Low, because these buyers already pay similar fees on prediction markets and option exchanges |
The table rests on three sources. The numbers in the How it pays and How often it trades rows come from the backtest: every weekly settlement of BTC and ETH from 3 January 2020 to 11 September 2026, 350 weeks for each, with LongVol's value tracked hour by hour through each week. The other rows are my reading of the payoff, which decides who each token suits, and of how traders use the closest products today, Deribit's weekly options and crypto prediction markets. Those options are traded, not just held to expiry. In the 24 hours to 13:45 UTC on Friday 9 October 2026, Deribit's 40 BTC options expiring a week later, on 16 October, traded 4,858 contracts of one BTC each, 33% of the 14,883 contracts open, and its 36 ETH options traded 20,167 of 90,538 contracts, or 22%.4 That was a busy day: from 1 to 10 October the same options traded 12.7% (BTC) and 9.2% (ETH) of that open interest on an average day, with BTC's ranging from 1.8% on the two Saturdays to 34.5% on Thursday 8 October.5 Over the seven days from Friday 2 to Thursday 8 October they traded 98% (BTC) and 68% (ETH) of it, close to one resale per contract a week, which is what the weekly flow below assumes for LongVol. LongVol gives holders the same reason to trade before Friday: in 35% of BTC weeks it doubled at some point, a median of 3.3 days after the open, a gain a holder can lock in by selling. Market makers would sit between the two sides, minting and redeeming pairs to keep supply balanced, so minting and redeeming stay free: a fee there would fall on them and come back as wider quotes.
Trading desks can also take the ShortVol side as a hedged trade. A pair is always worth the deposit, so holding ShortVol is the same as having sold LongVol at TRP's price. A desk that does this and buys the same payout on Deribit, by buying the 95% put and 105% call and selling the 80% put and 120% call, holds two positions that pay exactly opposite amounts at settlement, so it keeps the difference between the two prices whatever BTC does. Three costs come out of that difference: TRP's short-volatility fee, 1.75% × p × (1 − p) × deposit, a median 0.033% of BTC's price per pair; Deribit's exchange fee of 0.03% of BTC's price on each of the two options bought, 0.06%, because a combination pays only its dearer side;1 and half of Deribit's bid-ask spread on the four options, 0.055% of BTC's price and 0.127% of ETH's, measured on the 9 October 2026 chain.4 I priced the Deribit side each Friday from January 2020 to September 2026 at the implied volatilities fitted to that morning's trades between 07:00 and 09:00 UTC.6 Before costs, TRP's price was the higher one in 215 of the 350 BTC weeks and 276 of the 350 ETH weeks. After costs, the trade paid in 179 BTC weeks, 51% or about half, and in 253 ETH weeks, 72%. A desk sees both prices before trading, so it would join only in those weeks, adding ShortVol money when TRP's price is high.
An average week for $1M of BTC deposits
These are averages over the 141 BTC weeks in the backtest from 5 January 2024 to 11 September 2026, for $1,000,000 deposited as pairs. I assume buyers take all the LongVol at the Friday open and each LongVol is resold once, at a random hour of the week. Every number below is worked out from five averages over those weeks.
| Average over the 141 weeks | Value | What it is |
|---|---|---|
| p at the Friday open | 11.4336% | LongVol's price as a share of the deposit |
| p × (1 − p) at the open | 0.0958377 | Worked out week by week and then averaged, which puts it below 11.4336% × 88.5664% = 0.10126, because p changes from week to week |
| p during the week | 9.8774% | LongVol's value averaged over the 168 hours to the next open, which is what a resale at a random hour gets on average |
| p × (1 − p) during the week | 0.0741079 | The same average over those 168 hours, for the resale fee |
| LongVol's payout | 10.1239% | What LongVol received at settlement as a share of the deposit: nothing in 91 of the 141 weeks and the whole deposit in 3 |
Friday: the roll
During the week
Next Friday: settlement
Net result for the week
Fees are paid on top of each trade and never come out of the deposit, so the deposit always pays out in full. The resale moves $98,774 from one LongVol trader to another, so only its fees leave them, and what LongVol traders lose, ShortVol holders and TRP gain: $11,420 + $14,870 = $26,290.
Four ways to charge
I tested four fee schedules on the same weeks. For one pair at BTC $100,000, a calm week ends within 5% of the open, so ShortVol wins, and a breakout week moves more than 20%, so LongVol wins.
| One pair, one week | Calm week | Breakout week | ||
|---|---|---|---|---|
| LongVol pays | ShortVol pays | LongVol pays | ShortVol pays | |
| RiskON/RiskOFF fees2% a year on balances plus 15 bps a swap | $2.79 | $7.92 | $5.67 | $5.04 |
| Winners' rule3% of net winnings, proposed for Risk Prediction Markets, plus 15 bps a swap | $2.47 | $51.90 | $403.04 | $2.47 |
| 7% on both sides7% × p × (1 − p) × deposit on every trade | $102.68 | $102.68 | $102.68 | $102.68 |
| Recommended7% long-volatility, 1.75% short-volatility | $102.68 | $25.67 | $102.68 | $25.67 |
The recommended schedule charges the same every week, whoever wins, so both sides know the cost before they trade. The winners' rule charges mainly the winner: in a calm week (three weeks in five), ShortVol pays $49.43 on its winnings plus $2.47 on the swap, and in a breakout week LongVol pays $400.57 on its $13,352.17 of winnings, $403.04 including the swap fee.
| RiskON/RiskOFF2% plus 15 bps | Winners' rule3% of winnings | 7% on both sidesevery trade | Recommended7% and 1.75% |
|---|
The first five rows come from the 2024–26 BTC backtest, with one resale of each LongVol ticket and ShortVol's profit at TRP's prices. The last two are medians of the 400 Monte Carlo runs described below, where deposits come and go as users react to fees: deposits against the same run with no fees, and revenue against the best tested pair of rates.
- RiskON/RiskOFF feesSo small that in the Monte Carlo simulation the market keeps 96% of the deposits it would hold with no fees, against 81% under the recommended fees, but they collect less than a tenth as much.
- Winners' ruleSplits the cost evenly, but LongVol buyers pay only 2% of the ticket, a third of Kalshi's taker fee, so it raises about a third of the recommendation's revenue in the backtest and three fifths in the Monte Carlo simulation, while driving away almost as many deposits.
- 7% on both sidesEarns the most, but ShortVol then pays 5.9% of its premium every week, keeps under half its average profit, and in the Monte Carlo simulation half of all deposits leave.
- RecommendedPuts four fifths of the fee on LongVol, at the rate its buyers already pay elsewhere, and keeps ShortVol's yield worth having.
A fee on winnings has two more problems. It needs each holder's buying price, which tokens traded mid-week do not record, and it is easy to avoid: two hours before a calm week settles, a ShortVol holder owing $49.43 can buy a nearly worthless LongVol for under a cent and redeem the pair for $15,000, and LongVol can do the same in a breakout week. In the BTC backtest, that would have paid in 87% of calm weeks and saved 78% of the fee.
How fees change demand
The backtest, year by year
Revenue follows LongVol's price. On BTC it peaked in 2021, when LongVol cost 29% of the deposit on average, and was lowest in 2025–26, when it cost under 10%; ETH earns more because its LongVol costs more. Calm years still hold up, because the fee is a larger share of a cheaper ticket: 6.7% of the ticket when LongVol costs 4% of the deposit, against 5.6% when it costs 20%.
BTC
ETH
How the Monte Carlo simulation works
A backtest keeps the number of users fixed, but fees change who shows up. So I also ran a Monte Carlo simulation: one model of the market, stepped week by week through the same 350 weeks with users who react to fees, run 400 times with the users' reaction strengths drawn at random each time. Only the behaviour is random; every run uses the real prices.
- Real pricesThe same 350 weeks of BTC and ETH prices, priced with our risk engine.
- Users who reactShortVol money arrives after good weeks and leaves after heavy losses, LongVol buyers arrive after big moves and hedgers buy every week. Each side buys less as its fee rises.
- A price each FridayThe LongVol price settles where ShortVol's supply meets buyers' demand, and market makers absorb what is left.
- 400 random runsNobody knows how strongly real users react, so each run draws its 15 reaction settings at random from set ranges, such as how much less LongVol buyers buy as the fee rises.
Each run tests all 117 pairs of rates, 13 long-volatility rates from 0% to 20% times 9 short-volatility rates from 0% to 6%, and gives two numbers for each pair: deposits kept, the run's average deposits with the fees against the same run with no fees, and revenue, as a share of the most any pair earned in that run. The figures below are the median of the 400 runs for BTC, and ETH looks very similar.
What one more point of fee does
| Raising the rate | Revenue | Deposits kept |
|---|---|---|
| Long-volatility rate | ||
| From 6% to 7% | +8 points | −1 point |
| From 7% to 8% | +0.2 points | −6.5 points |
| Short-volatility rate | ||
| From 1% to 2% | +5 points | −6 points |
| From 2% to 3% | +5 points | −6 points |
Up to 7%, raising the long-volatility rate is almost free: each point adds about 8 to 9 points of revenue and drives away only about 1% of deposits. Above 7%, a buyer can get the same bet more cheaply on Kalshi or Polymarket (at 10%, the example ticket would pay $146.68, $44 more than either prediction market charges), so between 7% and 10% each point drives away about five times as many deposits and adds almost no revenue, and even at 20% revenue is only 9 to 11 points higher than at 7%, with deposits cut in half. This jump at 7% is built into the simulation, which assumes buyers compare TRP's fee with the prediction markets' fee, so launch data will have to confirm it. The short-volatility rate has no such jump: on BTC each point adds about 5 points of revenue and drives away about 6% of deposits, a steady trade-off. I set it at Kalshi's maker rate, 1.75%, where ShortVol still keeps nearly nine tenths of its average profit after fees.
Long-volatility rate, with ShortVol at 1.75%
Short-volatility rate, with LongVol at 7%
How sure I am
The reaction strengths are informed guesses, which is why the Monte Carlo simulation draws new ones at random in each of its 400 runs. Two parts of the model deserve care. First, revenue keeps rising above 7% in most runs, because hedgers and holders who roll their winnings barely react to fees, so the best tested pair usually sits at the edge of the tested range, and the revenue shares are measured against that edge.
Second, the simulation has no route to Deribit, so when LongVol demand is strong its Friday price clears above NTV, by a median of 23% on BTC and 6% on ETH. With the price held at or below NTV, the recommendation keeps about the same share of deposits, 80% on BTC, but earns 63% to 66% of the top revenue instead of 79%, and each point of the short-volatility rate adds under 1 point of revenue while driving away 8% to 9% of deposits, which makes a low short-volatility rate more important, not less. Only launch data will show how fast ShortVol money leaves when fees rise.
Try other rates
Kalshi and Polymarket charge takers 7%.
Kalshi's maker rate is 1.75%.
Each LongVol ticket is resold this many times a week.
Worked examples
Each example follows one pair through one week, from BTC at $100,000.
The LongVol buyer
The buyer pays $1,750.51 in all: $1,647.83 for the token and the $102.68 fee.
| BTC on Friday | Worth | Result |
|---|---|---|
| $103,000, up 3% | $0 | lose $1,750.51 |
| $92,000, down 8% | $3,000 | 1.7 times |
| $87,500, down 12.5% | $7,500 | 4.3 times |
| $78,000 or $124,000 | $15,000 | 8.6 times |
Without the fee, the best case would be 9.1 times. If BTC falls to $91,000 by Wednesday, the token is worth $4,168.24 with two days left, and selling it pays the $52.67 short-volatility fee and returns $4,115.57, which is 2.4 times the outlay.
The ShortVol holder
The token costs $13,352.17, the deposit less the $1,647.83 its LongVol sells for, and this week's roll costs $25.67, which is 1.56% of the premium.
| BTC on Friday | Worth | After the fee |
|---|---|---|
| $103,000, up 3% | $15,000 | gain $1,622.16 |
| $92,000, down 8% | $12,000 | lose $1,377.84 |
| $87,500, down 12.5% | $7,500 | lose $5,877.84 |
| $78,000, down 22% | $0 | lose $13,377.84 |
A calm week returns 12.1% after the fee, which is why ShortVol appeals to option sellers, and a move past 20% wipes out the token, which is why it is sized as a slice of a portfolio. At these prices every week, the fees would total $1,334.84 a year, 10% of the token's cost, against about $396 under the RiskON/RiskOFF fees and $3,120 in Deribit exchange fees for selling the same options each week.
What TRP earns from one pair
The resale adds the most because by Wednesday, with BTC at $91,000, p is 0.28 and p × (1 − p) is about twice as large.
How to launch it
Users pay on every trade on the Risk Marketplace, with a market maker or with each other, by hand or through the auto-roll. Designated market makers pay nothing, since they trade to keep the market running rather than to hold a position and a fee on them would widen every quote; minting and redeeming stay free for the same reason.
I would launch at 7% and 1.75%, and change a rate only if launch data shows one of two problems: weak LongVol demand or too little ShortVol money.
LongVol demand is weak
LongVol volume stays low for four weeks, and market makers are left holding LongVol that buyers do not take.
Lower the long-volatility rate towards 5%, a little above Deribit's exchange fee on BTC, which at typical prices works out to a rate of about 4.4%.
ShortVol deposits fall short
LongVol trades above NTV on the marketplace because ShortVol holders post too little money to meet demand.
Lower the short-volatility rate towards zero, which is what makers pay on Polymarket and on Kalshi's main BTC and ETH markets.
LongVol trading below NTV is not on its own a reason to cut, because Deribit already prices the same payout below NTV. A share of the yield earned on the idle USDC deposit and a discount for holders of a future governance token can both be added later, leaving the schedule as it is.
- Deribit, Fees: at the standard tier, options cost 0.03% of the underlying per contract, capped at 12.5% of the option price; a combination traded as one order pays only its dearer side; delivery costs 0.015%.
- Kalshi, Fee Schedule, effective 5 February 2026: taker fee round up(0.07 × C × P × (1 − P)); maker fee round up(0.0175 × C × P × (1 − P)) on the markets Kalshi designates for maker fees.
- Polymarket, Trading Fees, archived 10 October 2026: fee = C × feeRate × p × (1 − p), with a crypto taker rate of 0.07; makers pay nothing and share 20% of crypto taker fees as rebates.
- Deribit public API, book summary for all BTC and ETH options, 9 October 2026, 13:45 UTC.
- Deribit public API, hourly traded volume of the BTC and ETH options expiring 16 October 2026, from 1 to 10 October 2026.
- Deribit public API, BTC and ETH option trades made between 07:00 and 09:00 UTC each Friday from January 2020 to September 2026, used to fit each week's implied volatility.