Crypto Newbie / Simulators / CoW Swap Batch
CoW Swap Batch Auction Simulator
Most DEXs (Uniswap, Curve, Sushi) are first-come-first-served — your order is visible in the mempool, MEV bots front-run + sandwich it. CoW Swap is different: orders are SEALED, users sign INTENTS (not execution paths), and SOLVERS compete to find the best settlement. All winning trades clear at the SAME PRICE (Uniform Clearing). This simulator shows a batch of 6 sample orders, runs settlement (direct CoW matches first, then AMM routing), and computes solver economics across different competition levels.
Batch of orders (sample)
| Order # | Direction | Sell | Buy | Amount sold | Limit price | Status |
|---|---|---|---|---|---|---|
| #1 | sell | ETH | USDC | 5 | 3400 | ✓ Executed |
| #2 | buy | USDC | ETH | 18000 | 0.00028 | ✓ Executed |
| #3 | sell | WBTC | USDC | 0.5 | 92000 | ✓ Executed |
| #4 | buy | DAI | ETH | 7000 | 0.00028 | ✓ Executed |
| #5 | sell | ETH | USDC | 2 | 3380 | ✓ Executed |
| #6 | buy | USDC | WBTC | 45000 | 0.0000105 | ✓ Executed |
Market prices (USD)
Settlement result
Orders executed
7
Orders failed (limit not met)
0
Total user surplus (USD)
$368.00
Direct CoW matches
3
Saved by direct CoW (no AMM fee)
$249.00
Volume routed via AMM
$14,000.00
Solver economics
Winning solver reward
$366.16
Per-solver win rate
12.5%
Expected profit per attempt
$40.77
Annualised profit (24/7 operation)
$42,857,424
Why sealed-bid + batch matters for MEV protection
On Uniswap, your tx is broadcast to the public mempool. MEV bots see it, calculate the optimal sandwich, and front-run + back-run for profit. CoW Swap orders stay sealed until the batch executes (every ~30 seconds). Bots can't front-run an order they can't see. Even if a bot knew your order, they couldn't sandwich because all batch trades execute at the SAME CLEARING PRICE. No price impact from your specific trade = no MEV opportunity. This is structural MEV protection, not just trusting solvers to behave.
Coincidence of Wants — the 'CoW' in CoW Swap
When orders are batched, solvers look for direct matches: A wants ETH and is selling USDC; B wants USDC and is selling ETH. They can be matched INTERNALLY without touching any AMM. No fees, no slippage, no AMM spread (typically 0.3%). For batches with lots of opposing orders (common in volatile markets), 30-60% of volume can settle as direct CoW matches. The simulator's 'cowMatchesCount' shows this. The protocol launched CoW (the term) — it stands for Coincidence of Wants, the same economic concept as a barter exchange.
Uniform clearing price — fairness mechanism
All trades in a batch settle at the same price for the same pair. Why? Two reasons: (1) MEV protection — no within-batch ordering matters, eliminates front-running. (2) Fairness — all users get the same price, no advantage to whoever's signed first. This means a small order gets the same execution as a whale order (after slippage), which is fundamentally different from AMMs (where order size determines effective price). Sometimes the clearing price is BETTER than AMM (when solvers find good CoW matches), sometimes slightly worse (when batches are thin). On average, CoW Swap users get slightly better prices than direct AMM trades for large orders.
Solver competition — the protocol's incentive design
Solvers compete to win each batch. Winner = highest total user surplus (the solution that gives users the best aggregate price). Winner earns: a portion of the surplus (typically 10-50% after protocol fee) + transaction fee. Losers earn nothing (they paid gas to compete and got nothing). This is harsh — only ~1/N solvers per batch win, where N is the number of competing solvers. The simulator's 'expected profit per attempt' shows the harsh math: with 10 solvers each putting in 30s of work per batch, the average solver earns ~10% of the winner's reward. Top solvers (Barter, Naive, 1inch, Otex) dominate due to better algorithms and more efficient gas usage.
Frequently asked questions
+Is CoW Swap actually used much?
Yes — $20B+ lifetime volume, $50-100M daily volume in 2024. It's particularly popular for: (1) Large trades ($100k+) where MEV protection matters most. (2) Stablecoin swaps (perfect setting for CoW matches — high volume in both directions). (3) Multi-hop routes (CoW Swap can route through multiple protocols better than a single AMM). Most retail users don't know about it because the UX is slightly slower (30s batch delay), and the savings aren't visible to small traders. Whales love it.
+Who runs the solvers, and could they collude?
About 10-15 solver organizations compete on CoW Swap. Major ones: Barter, 1inch, Naive (the CoW DAO's own solver), Otex, Pyroshik. Solver collusion would be detectable: if they all submitted the same bad solution, users would see the price worse than what major AMMs offer at the same moment. Users could complain, CoW DAO could remove colluding solvers. Plus: it's hard to coordinate when entry/exit is permissionless. Still, the system has known centralization (a few solvers dominate volume); decentralization improvements are ongoing.
+What's the difference between CoW Swap and 1inch?
1inch is an AGGREGATOR — finds the best AMM route for your order and routes through them. Order is public, you can be front-run. CoW Swap is an AUCTION + RFQ system — your order is sealed, solvers compete to find the best path (which can include 1inch's route plus direct CoW matches). For most small swaps the two are similar. For large/sensitive trades, CoW Swap's MEV protection is meaningful. 1inch acquired Fusion in 2023 which is similar to CoW (sealed orders, solver competition) — convergence between the two models is happening.
+Can I run my own solver on CoW Swap?
Yes, permissionlessly. The solver SDK + API are open. To be competitive you need: (1) Excellent MIP/LP solvers (mixed-integer linear programming for batch optimization). (2) Real-time market data feeds. (3) Low-latency Ethereum nodes. (4) Capital for gas (you pay for failed attempts). (5) Solver bond requirement (1000 COW tokens staked, slashable for bad behavior). Realistically, only well-capitalized firms run profitably. Hobbyist solvers typically lose money to gas costs.
+How does CoW Swap fee compare to Uniswap?
CoW Swap fee: 0.1-0.2% (protocol fee + solver fee, much of which gets rebated to users). Uniswap fee: 0.05-1% depending on pool (V3 has 0.01%, 0.05%, 0.3%, 1% tiers). CoW Swap is competitive on cost — and CHEAPER once you factor in MEV savings (most users save 0.3-1% on average vs Uniswap due to no sandwiching). The biggest cost difference is GAS: CoW Swap user pays $0 in gas for failed orders (solver pays); Uniswap users pay gas regardless of success.