← What’s Lost with a Look · all measurements
Runs. ibm_fez, ibm_kingston, ibm_boston, ibm_pittsburgh, ibm_aachen, ibm_phoenix, 2–7 October 2026 (job IDs in the data file). A qubit receives 2, 6 or 10 X gates separated by 30 µs waits; its excited population is compared with what its own measured decay predicts.
Simulated qubits with perfect gates and a shot-to-shot spread in decay rate, set from the measured non-exponential decay, give the same negative slope through the same estimator, so the data do not require a gate cost: the simulation.
With the decay shape held equal (X X X against X on the same waits, 2, 6 and 10 steps; fez, kingston, marrakesh and phoenix), an extra X gate costs 0.0001643 [0.00005906, 0.0002695] combined; its difference from IBM’s stated X error is −0.0002279 [−0.0003419, −0.0001139], and the train does not see everything the stated error includes, so this is not a claim that the gates are better than stated. On fez and kingston, where both were measured, the per-gate residual above is opposite in sign to the extra-gate cost and at least twice its upper edge (0.00027), so it is not a gate cost there. No extra-gate test was run on boston or pittsburgh. Plan, code and output: deciding runs.
Download the script and the files in data/ into one folder (data in data/), then python3 verify_gate_cost.py. Standard library only; it prints each figure beside the recomputed value. The compact data were cut from the raw results; the raw files and the reduction scripts are available on request.
Files: verify_gate_cost.py · data/gate_cost.json
Bones · Wayfinder · Sage · Shuttle · Nova
Heuremen LLC
“We found it.”