The economic trilemma of onchain sports betting

9 min readOct 20


As onchain sports betting protocols mature, the topics of fees and driving value to different stakeholders emerge. This article looks at the difficulties that can arise when trying to keep all parties happy. A short glossary can be found at the end to clarify some terms if you are unfamiliar.

The most famous trilemma in the crypto space is the blockchain scalability trilemma. It suggests that a blockchain can only optimise for two of the following 3 characteristics: security, scalability, and decentralisation. For example, the bitcoin blockchain is secure and decentralised (if you don’t think too hard about the influence of large mining pools) but during times of high demand, fees and transaction times increase significantly. Arbitrum on the other hand, has low fees but is not decentralised as it runs off a single server. Some might say Solana solves this trilemma while opponents argue that its reliance on data centres, like Ethereum, means it is not decentralised, however this is not the point of discussion of this article.

You can’t please everyone. Pick a point in the triangle.

The sports betting protocol economic trilemma describes how protocol fee dynamics can only ever favour two of three economic participants in the system: bettors, liquidity providers (LPs), token holders. The objective of the fees of a protocol is to maximise the absolute return for token holders finding the maxima of the (volume * fee rate) vs fee rate; and for LPs by finding the maxima of (volume * margin) vs margin. However, these 2 functions do not exist in isolation, so really it is finding values of fee rate and margin which optimise for both parties at the same time. Another function to consider is volume vs token incentives, which is strictly increasing, but at the cost of token price, hurting token holders. Currently, sports betting protocols are still developing and the market is not efficient. They tend to operate in isolation without much focus on optimising and without considerations of the wider market.

Example functions of the volume — margin/fee and profit relationships discussed above.

Bettors’ only interest is getting the highest payout of any given bet. This means they want the highest odds and lowest protocol fees. This is something that a specific sector of traditional bookmakers offer: asian bookmakers. They offer a zero-thrills experience with few bonuses, offers, and exotic bets, and purely focus on tight spreads on core markets. All traditional bookmakers offer 0% fees beyond the margin built into the odds.

Liquidity providers’ only interest is getting fees on their LP position. This means they want either high margins or high volume from bettors, and want to avoid any profitable bettors — ideally every bet placed would lose. However, preventing profitable bettors from using an onchain protocol is not possible: even if they got their wallet blacklisted, it is trivial to make another address. Liquidity providers are also at the mercy of the core protocol working as intended and not offering wrong odds. In this context, LPs generally mean passive LPs like are used by Azuro, Overtime Markets, and, but the idea also applies to market makers on exchanges like SX Bet, BetDEX, and Aver.

Token holders’ only interest is value accrual back to the token. This is likely to be via token buybacks; dividend-like payouts; or utility functions for the token. The main utility which can be offered would be reduced fees or access to free bets and offers — both of which reduce the revenue which can be given to token holders.

A protocol can ignore token holders and support LPs and bettors by offering 0% additional fees beyond the spread in the odds. This means that the odds bettors see are the odds they get, while even a fee as small as 2% on profits will result in odds which are materially lower than expected and lower than competitors. For a passive LP-type protocol, it must then balance the margin in order to make providing liquidity attractive to LPs, but keep odds competitive for bettors. A low margin will attract more bettors as odds are better than competitors, while a high margin earns LPs more per dollar staked, but the volume will decrease as odds are unattractive. For active LPs (market makers) on exchanges, they can offer much more competitive odds if they are not charged a fee on their profits. Fees on profits mean the odds need to be worse to account for this, and then a fee for the takers makes their odds worse again, so it creates a negative feedback loop. If the exchange was not trying to reward token holders, these fees would be eliminated and market makers could quote higher odds and the volume would be much higher.

On the other hand, the protocol could favour bettors and token holders over LPs. This could happen in 2 ways: it could take a performance fee from LPs or avoid having LPs altogether. Taking a performance fee would be something as simple as taking a percentage of net LP profit each week, however this means that the yield on the LP position is directly reduced, and the protocol gets rewarded when LPs do well, but not when LPs do badly. Since LPs rely on the core protocol to correctly price bets and manage risk, this system makes it fairly unattractive to provide liquidity under these conditions. A protocol could avoid having passive LPs by entirely owning the liquidity pool. This could be done by selling the token in exchange for stablecoins which then make up the liquidity pool. There are some issues with this: token sales generally frowned upon these days; it involves selling lots of the protocol off at a low valuation early in the product lifecycle; limited liquidity or need to dilute early buyers to raise more. However, this setup aligns token holders and the liquidity pool since they are the same thing. Token holders earn their value accrual by the protocol successfully attracting bettors and volume, profiting from the margin in the odds, without having to charge additional fees to LPs or bettors.

Optimising for token holders and LPs over bettors is a rather pointless task, since bad effective odds means no volume so nothing to actually take the fees from. It might be possible to argue that fees can be offset by handing out protocol tokens to bettors, however if they are not aligned with the vision of the protocol they will just sell the tokens, decreasing the price. This means either the tokens are worth less next time round and the high fees and low token price make the odds unattractive; or the amount of tokens increase to be approximately the same dollar value, creating a negative feedback loop on the token value.

A look at existing protocols and where they sit on this trilemma plane is fairly difficult as most of the major protocols do not have a token yet.

Overtime Market which shares the THALES token with the options protocol only allows THALES token stakers to add to the liquidity pool. This drives value to the token but reduces the number of people willing to LP. Overtime has the margin and protocol fees (as the safebox fee) both built into the odds. The safebox fee was previously moved from 3% to 1%, which significantly boosted Overtime’s odds compared to competitors and favoured bettors over token holders. Most recently, the safebox fee was doubled to 2%; and 20% of weekly LP profit was sent to the safebox. With safebox fees being used to buy back and burn THALES tokens from the market, this represented a shift from favouring LPs to favouring token holders, while slightly reducing the odds offered to bettors.

A smaller, newer betting protocol called DEEPP launched with reasonable margins, no additional fees, token distribution to bettors, and the option for anyone to become an LP. After a few days, a fee of 2.5% was added to any winning bets. This shifted the odds from good/reasonable to poor for bettors, while favouring token holders as the fee was distributed to token stakers. Bettors could choose to see the token distribution as potentially making up for the fee, however the amount distributed is unpredictable, and selling the token can cause a negative feedback loop. As the fee reduces the competitiveness of the protocol for bettors, the volume is lower, so LPs earn less. This is an example of favouring token holders over bettors, and in turn, LPs.

SX Bet (an exchange) made a change this summer which involved reducing the taker fee from 4% to 2%, while keeping makers fees at 4% and discountable based on the number of SX tokens staked. This was a big boost to bettors, who are also rewarded SX tokens for betting. SX token stakers get fees distributed to them. While there was no direct improvement for market makers (basically LPs in this case), lower taker fees should have increased the volume from takers and so increased makers’ returns. Since the maker fee can be reduced by staking SX tokens and makers are more likely protocol power-users, they are more likely to see value in buying and staking SX tokens and are more likely to do it to reduce their fees in addition to receiving dividend-like payments. Overall, this taker fee reduction favoured bettors and hurt token holders directly. If volume increased as a result of this change, this would have helped LPs (who are the most likely token holders) and offset the fee reduction for token holders.

Recently, Azuro introduced promotional low margins on some leagues on the Gnosis blockchain. This is a boost to bettors who have access to some of the best odds in the world, however it hurts LPs who earn less for the risk they take in being the counterparty to these bets.

Purebet has no fees and the margin is determined by what is offered by other venues. Although there is a liquidity pool to facilitate betting, it is privately held and charges no additional margin. This is an example where the bettor is favoured entirely.

If a protocol had token-based governance, it is possible (and probably likely) that token holders would vote for a proposal which introduces or increases a fee, the proceeds of which are distributed to token holders or used to buy and burn the token. However, this proposal will come at the cost of hurting bettors, or LPs, or both. Any changes to the protocol and token price would be only in the short term, until the market adjusted to the new equilibrium which will result in LPs and bettors using other platforms which offer a better return on their risk. It could be prudent to ensure that LPs also have a vote in matters which affect liquidity provision, as they are more economically aligned with bettors. More importantly, protocols should ensure that token holders are a mixture of bettors and LPs, and only a small minority have no other economic involvement in the protocol. Striking the right balance of voting power between bettors and LPs will ensure the protocol reaches a fee-LP yield-odds equilibrium which is optimal.

Without removing one part of the trilemma, a protocol can only ever reach a compromise between all 3 parties. In the case of a protocol without token holders, there is still a balance to be struck between bettors getting good odds and LPs getting a high margin. The goal of the protocol team and governance is to use all the tools and leavers of margins and fees and token incentives and dividends and buybacks to reach an optimal equilibrium in a complex multivariable system which brings in the highest absolute return for each party in the trilemma. This equilibrium state will change over time as the wider market evolves and protocols are created and evolve.

Margin — the margin in the odds of an outcome, the vig, the overround. If an event is 50/50, odds of 2.0/2.0 is a 0% margin. If the odds are 1.9/1.9, that is a 5% margin.
Liquidity providers (LPs) — users who contribute to a pool of tokens which take in wagers and pay out winning bets. They are the counterparty to bets placed on non-exchange type protocols.
Exchange — a protocol which directly matches makers and takers on opposing sides of bets.
Token staking — some form of locking a token up so it is not on the open market. Usually locked for a set time, or can be unlocked on a time delay. Usually required to receive dividend-like rewards.
Safebox — specific to Overtime, a fund used for backstopping LPs in the event of a protocol error. Once over a certain amount, the excess is used to buy and burn THALES token.
Buyback — when a protocol buy’s its own token from the open market. Afterwards, it is usually either burned (removed from circulation) or recirculated into the economy in the form of incentives. In theory, this boosts token price over time as their is price-insensitive constant buying pressure.
Governance — the process of deciding on protocol parameters and actions, usually voted on by token holders.

About Purebet

Purebet is the first onchain sports betting aggregator. It aggregates odds and liquidity from 8 protocols across 6 blockchains ensuring Purebet users always get the best odds for their bet, and best execution for their size. Bets on placed at Purebet are settled seamlessly on Solana through our onchain exchange architecture and winnings are paid direct to wallets automatically.

Follow on Twitter:

Join our Discord:

Our website:




A decentralised sports betting exchange, liquidity aggregator, and degam ecosystem.