TITO Live: paper-trading results
Several versions of an intraday turning-point model trade 30 large U.S. stocks on live market data, side by side, with simulated money. Can any of them beat random trading and the market after paying real trading costs?
Hypothetical results. No real orders are placed and no money is at risk. Past or simulated performance does not predict future results. Nothing on this page is investment advice.
Today's scoreboard
Each version starts every day flat with a hypothetical $100,000. Net P&L is after paying the bid–ask spread and a fee allowance on every trade.
Net P&L, minute by minute
No P&L recorded for this day yet.
Prices and trades
▲ buy, ▼ sell, coloured by version. Hover a marker for details.
No prices recorded for this symbol and day.
Position, minute by minute (SPY)
Long, short or flat for each version. A change of colour is a buy or sell decision; no change is a hold.
Every trading day
No completed days yet.
Latest simulated trades
No trades yet for this day.
How the results are produced
A program streams live quotes from a brokerage data feed during U.S. market hours and builds one-minute price bars. After each minute closes, every version decides whether to buy, sell, exit or hold each stock. Decisions are filled immediately after the minute closes at the live quote: buys at the ask price, sells at the bid, plus a small fee allowance. So every trade pays the bid–ask spread. Open positions are valued at the midpoint between bid and ask, and every version is flat by 3:55 pm Eastern.
Each version has a hypothetical $100,000 and holds at most 20 positions of about $5,000 each. Two benchmarks run alongside: buying the S&P 500 ETF (SPY) for the session, and random trades at a similar rate. A version is only interesting if it beats both after costs, over many months: a few days or weeks of results are mostly luck.
The versions are variants of the TITO (Turning-point Inference and Transition Observer) research program by Thomas Rhee and Mark Keith. Their rules and parameters are not published here; the research write-ups report the evaluation methods. Versions are added only after they pass historical tests registered in advance.