Ask-heavy order books predict winning longs 66% of the time across 329 trades — but the short mirror fails

By Strota Newsroom · 2026-08-20 · How Strota reports

Can you tell which intraday long trades will win just by reading the order book? Strota scanned 329 resolved trades across six sessions from June 9 to June 16, 2026, looking for markers that lift a trade's win rate above the 57% baseline. The strongest single marker turned out to be the ask-heavy order book — a state where resting sell orders outnumber resting buy orders above the stock's current price. In plain terms, more sellers are standing in the way than buyers are propping it up. Ask-heavy setups won 66% of the time, nine percentage points above baseline, and they did so on five out of six days.

Why does a 'bearish' book state produce winning longs? The mechanism is absorption. When a stock breaks its intraday high while the order book is ask-heavy, aggressive buyers are eating through visible supply. That demand is real and measurable. The directional split confirms it: ask-heavy on longs hit 74% win rate (76 trades), while ask-heavy on shorts came in at just 57% — essentially baseline. The mirror state, bid-heavy (resting buyers below the price), produced the worst long performance at 46%. Passive buyers sitting below the price suggest no urgency; the stock drifts down into them rather than being pushed up through them.

The magnitude of the order-book imbalance did not matter as a gradient. Setups with 25–50% imbalance won 67% of the time, but those at 50–75% actually performed worse at 59%. Ask-heavy is a binary present-or-absent signal — either the book is ask-heavy or it is not. Weighting the degree of imbalance adds noise rather than information.

Stacking ask-heavy with other confirmers pushed win rates higher. Ask-heavy plus conviction-4 (a specific signal-strength threshold) won 78% of the time across 55 trades and was profitable on every single day in the sample. Ask-heavy combined with above-VWAP (a price remaining above the volume-weighted average price for the session) hit 76%. When all four confirmers aligned — ask-heavy, with-trend long, at the day's high, and above-VWAP — the win rate reached 80% across 30 trades, positive every day. Removing any one element dropped the rate, with gap absence being the most damaging cut: ask-heavy plus no gap fell to 62%.

Several counter-intuitive findings emerged. Volume blow-offs did not help: setups with no volume spike won 70% versus just 53% when volume exceeded fifteen times normal. Conviction scores were non-linear — the sweet spot was conviction-4 at 61%, while higher scores at 5 and 6 actually degraded to 53%. Neither volume magnitude nor conviction level should be treated as a simple more-is-better input.

The short side told a different story. The hypothesised mirror — bid-heavy plus below-VWAP plus day's low — barely appeared. Bid-heavy showed up in just 4 of 94 shorts, and those four won only 25% of the time. Stocks breaking their daily lows carried ask-heavy books (70 of 94 trades), meaning the order book was confirming selling pressure rather than signalling absorption in the other direction. Shorts overall won 52% and were positive on only two of four days. The absorption mechanism is long-only; it does not reverse for shorts.

Hard gates emerged from the loser markers. IGNITING-stage entries won just 27%, gaps of 3% or more won 33%, bid-heavy longs won 46%, and flat setups with less than 2% extension and no gap won 49%. All of these are below the baseline and represent setups to avoid.

The limitations are real. Six sessions is a small sample, and the period was bullish-skewed. One day — June 12 — accounted for 59% of all profit, so the average R figures are inflated by a single trend day. The in-sample 80% A-plus figure rests on just 30 trades. The combinations that held on every day (ask-heavy plus conv-4 on 6 of 6 days, the above-VWAP stack on 5 of 5) are the more trustworthy findings. Everything else is a hypothesis waiting for walk-forward validation. This is a statement about the past, not a forecast.

More money stories

• Tested on fresh data, our high-win-rate setup wins less often but pays more
• Costs, not gross returns, decided which trades to skip: a sector filter earned more on fewer trades
• Big gaps don't sustain: Indian stocks that open far from the last close mostly hand it back
• Volume was the only thread in our biggest movers, and it was unknowable at entry
• Buying weakness worked, buying strength lost: results from a broad signal sweep
• An 80% win rate collapsed to 31% the first time we tested it forward