The engine behind my systematic bitcoin book is simple to describe. Every day at the close, it produces three forecasts for bitcoin: where price is likely headed 9 days out, 20 days out, and 60 days out — call them the short, medium, and long view. When bitcoin is trading below its recent average and all three point up, with conviction building the further out you look, the model buys. When the short and medium views roll over and turn negative, it sells short. It's impatient on the way up and patient on the way down: it takes profit on longs the moment momentum fades, but holds shorts until the trend clearly repairs. Downtrends die slower than rallies, and the book is built to respect that.
Here's why I trust it — this is the same model that called the top. In October, with the market euphoric near $124,000, its medium-term view collapsed and it went short, essentially at the high of the cycle. It rode bitcoin down toward $65,000, and it's still short today, in profit, from around $76,000. When its exit conditions finally trigger, the book flips to the buy side and starts accumulating the floor.
The track record — run on real bitcoin price history, net of trading costs — tells the story. Over the last two months the model is up about 14% while simply holding bitcoin lost 14%. Over six months, the stretch that captured the whole slide from the highs toward $65,000, it returned roughly 67% while holders lost 39%. That's the edge: it makes money when the market falls, and compounds while everyone else waits to get back to even.
Now the honest part, because you should hear it from me: the last 30 days were negative. Bitcoin chopped higher off the lows while the book stayed short, and the model gave back about 7% against a market that drifted up 2%. It has losing months — it just pays for them many times over across the longer horizons. If a bad month will shake you out, this isn't your trade.
So here's the play for the group. We market-make the $60,000 to $90,000 band with this model as the brain. The book runs on Base against Coinbase's wrapped bitcoin — one-to-one custodied — and it's skewed short right now from that $76,000 entry, harvesting toward the floor and flipping long the moment the model turns. Your liquidity backs the book through that rotation. The range pays the spread; the model decides the side.
A Yes signs you into the market-making position with your own wallet. A No costs you nothing. Either way, every fill lands on-chain for the group to see.
Past performance doesn't oblige the future to cooperate — but this is a machine with a documented record, not a hunch.
Prototype / research scenario — not investment advice.