▤Stale quotes in crypto market data: how often they happen, and what they do to a backtestData quality · Measured Aug 12, 2026 · one full trading day (20:00 to 20:00 ET)STALLED MINUTES8.5-19.0%LONGEST STALL1,441 min›▤The real cost of crypto trading: spreads, order book depth, and what your order actually paysExecution · Measured Aug 12, 2026 · one full trading day (20:00 to 20:00 ET)UNFILLABLE AT $25K3-65%TIGHTEST QUOTE3.0 bpDEPTH SPREAD27x›▤Loading a day of tick data: 73.8s to 0.9sPerformance · Measured Jul 28, 2026 · recordingLOAD TIME73.8s → 0.9sSYMBOLS222RECORDING1.59 GB›▤Why Kraken ticker prices can appear stale on quiet marketsMarket Data · Measured Aug 9, 2026 · single live sessionLONGEST STALL OBSERVED436 s›▤How we verify two backtest engines produce the same resultsEngine Parity · Measured Jun 29 – Jul 7, 2026 · five recorded sessionsVALUES COMPARED3,373,795MISMATCHES0MODULES17›▤One tick of slippage erased 44% of the backtest edgeExecution · Measured Apr 15 – Jul 31, 2026 · 75 sessionsEDGE REMOVED, 1 TICK44%TWO TICKS87%TRADES2,084›▤Does a faster stock screener catch more movers? We tested 501 sessionsMarket Discovery · Measured Aug 2, 2024 – Aug 6, 2026 · 501 trading sessionsCATCH RATE, 5S–300S88.4–90.6%MOVER EVENTS75,957SESSIONS501›▤82% of our backtest fills were trades we could not have taken liveExecution · Measured Apr 15 – Jul 31, 2026 · audit over 564,303 eligible candidate rowsFILLS INELIGIBLE82.2%PROFIT IN THOSE FILLS48%›
Research: measured trading data
Measured tests from TraderWe — execution, slippage, data quality and backtest realism, with the method and the numbers.