ATS-1

Performance

your account vs. just holding BTC refreshes every 5s

No trades closed on this run yet. The system polls a new 1-hour bar every 30s — if a strategy is gated by the current regime (range), it sits and watches until the regime flips.

Your account

$10,000.00

Just holding BTC

$10,105.64

Difference

−$105.64

Trades closed

EV $0.00 · 95% CI [0.0000, 0.0000] · n 0

Best strategy

none

EV $0.00 · CI [0.0000, 0.0000] · n 0

Active positions

$0.00 deployed

Equity

last 30 days refreshes every 30s

Only 1 equity point recorded so far — a line needs at least 2. The next point gets written when the coordinator processes the next completed 1h bar.

Strategies

how each one's doing refreshes every 5s

No strategy rows yet for this run — strategies appear here once they take their first trade or get their first capital allocation.

Activity

last 20 events refreshes every 5s

Quiet so far — no trades, no signals, no regime flips yet.

Mutations

parameter changes in flight refreshes every 30s

Shadow-validating

Nothing in shadow validation right now.

Applied this week

No mutations applied this week.

Rolled back

No rolled-back mutations.

What do these terms mean?

Expected Value (EV): The average profit or loss per trade. P(win) × avg_win + P(loss) × avg_loss. Positive EV means each trade is, on average, profitable. EV beats raw win rate because a strategy can win 90% of trades and still lose money if the losing 10% are huge.
95% CI: A 95% confidence interval on EV. The true average per-trade outcome is likely inside this band. Wide CI = uncertainty (few trades). Narrow CI = the number is trustworthy.
Win rate: Fraction of trades that closed profitable. Never shown alone — always paired with EV and CI, because it can mislead on its own.
Regime: How the classifier sees the market right now. One of panic, euphoria, squeeze, trend, or range. Each strategy is only active in regimes it's allowed to trade.
Probing size: The smallest position size the risk manager will allocate. Used when a strategy has fewer than 30 trades — not enough data to size with confidence. Strategies in this state show "still learning".
Mutation: A proposed change to a strategy's parameters, generated by the meta-learner. Every mutation is shadow-validated against historical data before going live.
Shadow validation: Running a proposed mutation against past data in an isolated in-memory database. If it doesn't pass statistical gates there, it never touches the live strategy.
Buy & hold: The benchmark. What your account would be worth if you'd just bought BTC at the start of the run and held. ATS-1's job is to beat this line.