Buy filters: helping or costing?
Setups stopped by just one filter get a practice trade too. If they do worse than the setups the bot buys, the filter is helping.
Trial buy rules (on paper only)
Stricter buy rules being tested. They never stop a buy: each buy and practice trade is only marked with the ones it would fail.
Win rate by trait
Practice trades of setups that passed every filter, graded with the bot's real sell rules. Weight = how much learning moves the buy score (capped at ±0.05 each); that column shows up once learning has 20 results.
Real trades
Closed positions, by who bought them and why they sold. "+ your add-on buy" = you added shares with Buy to lower the average.
Order costs
Every order. vs mid = what was paid above (buys) or below (sells) the middle of bid and ask; + = better. Missed = not filled, e.g. a mid-price buy skipped because the price ran up. Spread = average bid-ask gap.
Latest results