In DeFi market making, how much does it really cost to quote 11 seconds too slowly?
A study of 4 real USDC-WETH liquidity pools came up with a striking figure: 84.5%.
The study compared market makers that updated their quotes every 12 seconds with venues that updated them every second.
The results showed that increasing the quote frequency from once every 12 seconds to once every second could eliminate about 84.5% of expected price impact.
Why is the gap so large?
Suppose ETH is rising rapidly. A venue updating quotes every second can quickly reflect the new price; but an AMM that updates more slowly may still be selling ETH at the old price.
Arbitrageurs can profit from this time lag, while market makers bear the losses.
This is a form of adverse selection: your trading counterpart may have access to information about price movements before you do.
In theory, slow market makers can screen trades and reject orders more likely to come from traders with the latest price information. But the study estimates that, to break even with venues updating quotes every second, market makers quoting every 12 seconds would need to filter out about 84.5% of informed order flow.
This refers to informed flow, not all orders.
More notably, a simple Gaussian model yielded only 71.1%, below the threshold implied by data from real pools. This suggests that underestimating the complexity of market price movements could lead to overly optimistic market-making strategies.
Of course, these results come from a specific model and a sample of 4 pools. Actual screening ability, costs, and other market conditions can also affect the results, so they should not be applied directly to all AMMs.
But for AMMs, LPs, and RFQ market makers, the core issue is clear:
How much you earn in fees is one thing; how much faster traders take from you when your quotes go stale is another.
Market making isn't just about calculating revenue—you also have to account for the cost of being exposed to an information disadvantage.
A study of 4 real USDC-WETH liquidity pools came up with a striking figure: 84.5%.
The study compared market makers that updated their quotes every 12 seconds with venues that updated them every second.
The results showed that increasing the quote frequency from once every 12 seconds to once every second could eliminate about 84.5% of expected price impact.
Why is the gap so large?
Suppose ETH is rising rapidly. A venue updating quotes every second can quickly reflect the new price; but an AMM that updates more slowly may still be selling ETH at the old price.
Arbitrageurs can profit from this time lag, while market makers bear the losses.
This is a form of adverse selection: your trading counterpart may have access to information about price movements before you do.
In theory, slow market makers can screen trades and reject orders more likely to come from traders with the latest price information. But the study estimates that, to break even with venues updating quotes every second, market makers quoting every 12 seconds would need to filter out about 84.5% of informed order flow.
This refers to informed flow, not all orders.
More notably, a simple Gaussian model yielded only 71.1%, below the threshold implied by data from real pools. This suggests that underestimating the complexity of market price movements could lead to overly optimistic market-making strategies.
Of course, these results come from a specific model and a sample of 4 pools. Actual screening ability, costs, and other market conditions can also affect the results, so they should not be applied directly to all AMMs.
But for AMMs, LPs, and RFQ market makers, the core issue is clear:
How much you earn in fees is one thing; how much faster traders take from you when your quotes go stale is another.
Market making isn't just about calculating revenue—you also have to account for the cost of being exposed to an information disadvantage.