Latency Models in HFT Backtests Explained
Trade prints per second through one full session, ET clockseries ·
2026-09-24 · 78×2
How long a top of book quote survives, by latency budgetranking ·
2026-09-24 · 9×2
Average distance the mid price travels over each latency horizonranking ·
2026-09-24 · 4×2
Share of quotes replaced within 5 ms and 50 ms, by symbolranking ·
2026-09-24 · 5×3
Trade prints per second through one full session, ET clock
Trade prints per second through one full session, ET clock
| et_time | trades_per_second |
|---|---|
| 09:30 | 109.5 |
| 09:35 | 48.6 |
| 09:40 | 37.5 |
| 09:45 | 38.7 |
| 09:50 | 31.7 |
| 09:55 | 37.2 |
| 10:00 | 45.4 |
| 10:05 | 42.2 |
| 10:10 | 41.1 |
| 10:15 | 36.9 |
| 10:20 | 41.2 |
| 10:25 | 38.7 |
| 10:30 | 44.7 |
| 10:35 | 39.4 |
| 10:40 | 45.6 |
| 10:45 | 35.7 |
| 10:50 | 38.9 |
| 10:55 | 44.6 |
| 11:00 | 38.2 |
| 11:05 | 36.6 |
| 11:10 | 39.8 |
| 11:15 | 37 |
| 11:20 | 36.4 |
| 11:25 | 34.7 |
| 11:30 | 34.9 |
| 11:35 | 32.3 |
| 11:40 | 34.8 |
| 11:45 | 31.6 |
| 11:50 | 36.3 |
| 11:55 | 33.9 |
| 12:00 | 31.6 |
| 12:05 | 35.8 |
| 12:10 | 31.6 |
| 12:15 | 16.2 |
| 12:20 | 21.6 |
| 12:25 | 16.4 |
| 12:30 | 17.1 |
| 12:35 | 19.1 |
| 12:40 | 19.1 |
| 12:45 | 17.7 |
| 12:50 | 17.2 |
| 12:55 | 17.8 |
| 13:00 | 14.3 |
| 13:05 | 14.2 |
| 13:10 | 11.8 |
| 13:15 | 10.7 |
| 13:20 | 11.5 |
| 13:25 | 12.8 |
| 13:30 | 15.1 |
| 13:35 | 12.5 |
| 13:40 | 11.8 |
| 13:45 | 14.5 |
| 13:50 | 12.6 |
| 13:55 | 18.5 |
| 14:00 | 15.2 |
| 14:05 | 15 |
| 14:10 | 14.3 |
| 14:15 | 14.3 |
| 14:20 | 14 |
| 14:25 | 15.1 |
| 14:30 | 12.8 |
| 14:35 | 14.9 |
| 14:40 | 16.8 |
| 14:45 | 15 |
| 14:50 | 14.9 |
| 14:55 | 16.3 |
| 15:00 | 14.7 |
| 15:05 | 14.2 |
| 15:10 | 15.3 |
| 15:15 | 20.7 |
| 15:20 | 18.1 |
| 15:25 | 18.3 |
| 15:30 | 22.9 |
| 15:35 | 21.7 |
| 15:40 | 25.6 |
| 15:45 | 29.3 |
| 15:50 | 72.7 |
| 15:55 | 112 |
the exact SQL behind every number
SELECT
formatDateTime(toStartOfFiveMinute(toTimeZone(window_start, 'America/New_York')), '%H:%i') AS et_time,
round(sum(transactions) / 300, 1) AS trades_per_second
FROM global_markets.delayed_stocks_minute_aggs
WHERE ticker = 'AAPL'
AND window_start >= '2026-09-15 13:30:00'
AND window_start < '2026-09-15 20:00:00'
GROUP BY et_time
ORDER BY et_time
More from this analysisLatency Models in HFT Backtests Explained
How long a top of book quote survives, by latency budget
ranking 9×2
→
Share of quotes replaced within 5 ms and 50 ms, by symbol
ranking 5×3
→
Average distance the mid price travels over each latency horizon
ranking 4×2
→
The best bid and the price levels it visited, AAPL, 15 minute buckets
series 24×3
→
See all 3,293 queries →