Execution gap power
Execution gap power is one of 21 measurements this engine publishes in full. It is regenerated from a real run rather than transcribed, and it is shown here with its own claim boundary so the number and its limits arrive together.
The measurement
- Any gap detectable today
- no
- Go live date
- 2026-08-07
- Schema
- canli.alphac-execution-gap-power.v1
- Verdict
- NOT MEASURABLE YET — publish the power, not a number
Sleeves
Nested beyond what this page renders. Read it in the artifact.
The answer. THE RECORD IS TOO SHORT. Overlap between the live marks and a walk-forward that extends into the live period exists for two sleeves and is 5 and 9 days — four and eight return observations. No execution gap is distinguishable from noise at that length, and the observed divergences below are OBSERVATIONS, not estimates.
The quantity that answers it soonest. The TRACKING DIFFERENCE d = r_live - r_backtest, not either Sharpe. The two run the same signal on the same days, so their returns are strongly correlated and the difference series is far less noisy than either level. The record needed to detect an execution gap is therefore much shorter than the record needed to establish the edge itself — which is the useful thing to know.
What would change this. Time, and nothing else. The days_required figures below are the honest answer to 'when can we ask'. They also depend on the tracking sd being estimated from four and eight observations, which is itself a very noisy estimate — treat them as an order of magnitude, not a date.
When to look again
A 1pct annual gap
- AlphaMax
- 6.7 years
- AlphaTrend
- 9.2 years
A 5pct annual gap
- AlphaMax
- 3 months
- AlphaTrend
- 4 months
Note. The useful output is not a number today but a date. Converted from return observations at 252 trading days a year, and rounded outward because the tracking sd behind them is itself estimated from four and eight observations.
The practical reading. A LARGE execution gap — 5% of equity a year — becomes visible in roughly three to four months, so that is when this is worth re-running. A 1% gap needs the better part of a decade and is not answerable by waiting. If a gap that small matters, it has to be attacked by measuring fills directly against their decision prices, not by comparing equity curves.
Why only two sleeves. A comparison needs a backtest that covers the live days. The BLESSED research curves (crypto_carry_wk, k30_dn_63) are frozen by disclosure protocol and end 2026-06-01, before go-live, so they cannot answer this question at any record length. Only the live-forward walk-forwards, regenerated by the daily ticks, overlap at all.
Check it yourself
Every figure above is read from /glassbox/execution_gap_power.json, the
artifact the engine wrote. Nothing on this page is typed by hand: the page is generated from that
file, so a figure that moves in the artifact moves here and a figure that is not in the artifact
cannot appear here at all.