CMS Holdings

Relative value and financing trades. When a sector moves, the desk buys the smaller names around it and waits for them to catch up. If a contract to buy bitcoin in three months costs more than bitcoin does today, the desk sells that contract and buys the coin. The two prices have to meet on the day the contract settles, so the difference between them is the profit, whichever way the market moves in between.

Every position is scored by how hard it moves with bitcoin, and all of them count against one risk limit, measured both in bitcoin and in dollars. When the book is already at that limit, he cannot simply add to it: buying something new means cutting or hedging a position he already holds.

In a liquidation cascade the desk buys what is being dumped and sells into the bounce, but checks the spot market first: if real coin is being sold there, the fall is not a mistake worth catching. The machines only fill the orders; what to own, and how much, is decided by hand.

The funding rate, the fee longs pay shorts to hold a perp, used to tell him when the market had got too long. That stopped working once perps spread beyond BitMEX, so he reads open interest in the alts instead. He also keeps a list of the year's best performing tokens as a gauge of appetite: while those are still being bought on every dip, money is still coming into the market, and when the buying stops, people are getting out. Crowding is the risk he keeps naming: when every liquid fund owns the same basket, the edge is gone and they all reach for the same exit.

He learned to trade at a hedge fund run by a former Chicago options pit trader. He holds losers to zero out of spite and sells them the first time they come back. Traders from traditional markets are too rational for crypto, taking profit long before a move is finished, and his advice to everyone else is to sit on your hands.