Median
Not available
Records
0
What users actually receive.
We compare completed destination payouts with the strongest available reference for each provider: an independently reconstructed official quote or an explorer-reported expected output. Current results are preliminary and expand as verified transactions are added.
Conclusion
Our first evidence shows that a live quote and a completed payout are not the same measurement. Harbor's explorer expected output was lower than final payout in every curated record, while Chainflip's explorer also reported negative Swap Price deltas. These signals justify further testing, but they do not yet prove that either frontend quote was inaccurate.
Observed performance by provider
Each row summarizes the latest collected evidence. The default view uses verified collection sessions from the 30-day research window; it does not imply continuous probing. Selecting a benchmark size narrows the table to records in that category. The evidence basis remains visible because reconstructed quotes and explorer expected outputs are not exact accepted quotes.
Filters
Median
-54.0 bps
Records
5
Median
-16.6 bps
Records
11
Median
Not available
Records
0
ProviderEvidence basisRecordsMedian
THORChainOfficial quote probes and protocol executionsReconstructed quote curvesIndependent quote curves compared with protocol outcomes0Not availableChainflipswap.chainflip.io onlyReconstructed quote curvesOfficial main broker quote curves compared with public outcomes5-54.0 bpsHarborHarbor explorer recordsExplorer expected outputExplorer expected output compared with final destination payout11-16.6 bpsNEAR IntentsOfficial 1Click quotes and public intent executionsReconstructed quote curvesIndependent dry quote curves compared with public outcomes0Not availableBenchmark sizes
Choose a size to update the provider comparison above.
All benchmark sizes
Evidence levels
These levels describe how closely quote evidence surrounds an execution inside a collection session; they do not imply that the site probes continuously. Exact accepted quotes remain separate from reconstructed curves. Only exact matches and two-sided curves within five minutes may be eligible for cautious aggregate quote accuracy analysis, and provider and outcome filters still apply.
1
Exact accepted quote
A captured accepted quote linked to the same completed execution. This is not currently available because CheckYourSwap does not execute the user's swap or receive its accepted quote identifier.
2
Two-sided reconstruction
Independent quote curves before and after the execution reference time, with each curve no more than five minutes away. This is the strongest currently available reconstructed evidence.
3
Supporting time-aligned reconstruction
One curve within five minutes, or curves on both sides within fifteen minutes. Useful as supporting evidence, but excluded from the primary quote-accuracy conclusion.
4
Wider supporting curve
The nearest usable reconstructed curve is within one hour but does not meet the tighter timing levels above. It is retained for research and excluded from quote-accuracy conclusions.
Guaranteed minimum measurement
Where the transaction contains a minimum, we compare it with the executed swap output as a separate price-protection check. For Chainflip, the fill-or-kill minimum is measured before the destination egress fee. THORChain streaming-parent memo targets are per-subswap limit prices, so they are excluded from total-output guarantee counts. This measurement does not prove that the headline quote was accurate.