What Actually Happens When You Trade on an AMM

Most people learn about AMMs from a whitepaper or a lecture and walk away thinking they understand how it works. They don't. Not really. Here is what happens when you try to use one, based on actual experience. An Automated Market Maker replaces the traditional order book with a smart contract that holds liquidity pools. Instead of matching a buyer to a seller, you trade against a pool of tokens. The price is determined by a formula — most commonly a constant product formula where the product of the two token reserves stays the same before and after a trade. You put token A in, you get token B out, and the ratio shifts. The math is simple. The consequences of that simplicity are not.

The constant product formula is x * y = k. If a pool has 1000 ETH and 3,000,000 DAI (so k equals 3,000,000), and someone swaps 1 ETH into that pool, the new ETH reserve is 1001. To keep k constant, the DAI reserve drops to about 2,997,003. You receive roughly 2,997 DAI. That seems fine until you try to swap 100 ETH. The math breaks your expectations fast. Step one: pick a protocol with audited code and a track record. Uniswap v3, Curve, and Balancer are the most widely used. Each has different strengths. Curve dominates stablecoin trading. Uniswap handles broader token coverage. Balancer allows custom weightings. Choose based on what you are actually trying to do, not what sounds impressive. Step two: calculate your maximum acceptable slippage before you enter any trade. For stablecoin pairs, 0.05% is usually fine. For volatile pairs, 0.5% to 1% is more realistic. Anything higher means the pool is too thin for your trade size.

Step three: if providing liquidity, start with a wider price range or stick to concentrated liquidity only if you are prepared to rebalance regularly. I rebalance my concentrated positions every 7 to 14 days depending on market volatility. Missing a rebalance window for more than three weeks has cost me noticeable fee income in the past. Step four: track your impermanent loss separately from your fee earnings. Most portfolio trackers do not do this automatically. You need to know whether your fees are actually covering the IL or whether you are slowly losing ground. I use a simple spreadsheet for this because no tool I found was reliable enough to trust blindly.

The bottom line on Amm Chapter 20 style material is that the theory is straightforward. The practice is where it gets complicated. The formulas do not lie, but they also do not tell you everything you need to know. Real experience with slippage, range management, and market conditions teaches you what reading about it never will.