Evidence receipt: Assay 002

Receipt status

Published with frozen evidence.

The receipt binds the published result to the sealed export and its conformance record. It does not regenerate any trading figure.

Receipt: assay-002-lorentzian-classification-btc-eth-v1
Markets: BTC/USD and ETH/USD
Pairs: 2
Result: 2 falsified
Evidence state: Sealed
Publication: 8 July 2026

Identity

What the receipt binds.

Target class: Category-level open-source mechanism; no named implementation, vendor, product or person
Mechanism: Lorentzian-distance k-nearest-neighbours classification
Evaluation window: 1 January 2025 to 30 June 2026
Warmup history: From 1 January 2023
Markets: BTC/USD and ETH/USD
Cost model: 84.0 bps round trip
Evidence engine commit: 50df0c8
Conformance record: assay-002-lorentzian-conformance.md

Result summary

Two pairs. Both claims failed.

2frozen pairs

2negative accounts

84 bpsround-trip cost

328round trips

Pair	Starting account	Ending account	Return	Buy-and-hold ending	B&H return	Round trips
BTC/USD	$10,000.00	$2,608.67	−73.91%	$6,184.76	−38.15%	167
ETH/USD	$10,000.00	$1,848.04	−81.52%	$4,625.78	−53.74%	161

Both pairs fail both falsifier legs: the worked account finishes below the starting account and below same-window buy-and-hold after the same cost treatment.

Public files

Inspect the sealed record.

Open the article, measured account bridge and public verification records attached to this receipt.

Read Assay 002Return to the full reader article Account bridge JSONMeasured gross, net and cost series Bridge receiptDerivation identity and reconciliation Bridge checksumsPublic verification record

Configuration

The frozen test frame.

Direction: Long-only
Neighbours: 8
Label and exit horizon: 4 bars
Execution timing: Confirmed-bar signal, next-bar open fill
History window: 2,000 bars
Normalisation: Expanding min-max to minus one and plus one
Filters: Volatility, regime and kernel enabled; ADX disabled
Cost treatment: Static Kraken taker assumption, no depth-walk

The exported capsule retains README.md as its method note. The public article repeats the tested mechanism, window, account treatment, benchmark and cost model.

Show exact exported configuration

Field	Value
`classifier_config.adx_filter_enabled`	`False`
`classifier_config.adx_period`	`20`
`classifier_config.adx_threshold`	`20.0`
`classifier_config.assay_receipt_id`	`assay-002-lorentzian-classification-btc-eth-v1`
`classifier_config.cci_period`	`20`
`classifier_config.confirmed_bar_signal`	`True`
`classifier_config.distance_metric`	`sum_log1p_absolute_feature_delta`
`classifier_config.dynamic_exits_enabled`	`False`
`classifier_config.fixed_exit_bars`	`4`
`classifier_config.history_window_bars`	`2000`
`classifier_config.implementation_scope`	`keel_native_no_third_party_trading_script_code`
`classifier_config.kernel_crossover_lag`	`2`
`classifier_config.kernel_filter_enabled`	`True`
`classifier_config.kernel_lookback`	`8`
`classifier_config.kernel_regression_level`	`25`
`classifier_config.kernel_relative_weight`	`8.0`
`classifier_config.kernel_smoothing_enabled`	`False`
`classifier_config.label_lookahead_bars`	`4`
`classifier_config.long_only`	`True`
`classifier_config.neighbors`	`8`
`classifier_config.next_bar_open_fill`	`True`
`classifier_config.normalisation`	`expanding_minmax_to_minus_one_plus_one`
`classifier_config.regime_filter_enabled`	`True`
`classifier_config.regime_threshold`	`-0.1`
`classifier_config.rsi_period`	`14`
`classifier_config.second_rsi_period`	`9`
`classifier_config.sparse_sampling`	`same_four_bar_phase_with_strictly_prior_labels`
`classifier_config.sparse_step_bars`	`4`
`classifier_config.volatility_filter_enabled`	`True`
`classifier_config.volatility_long_atr_bars`	`10`
`classifier_config.volatility_short_atr_bars`	`1`
`classifier_config.wave_trend_average_length`	`11`
`classifier_config.wave_trend_channel_length`	`10`
`cost_model.cost_model_static`	`True`
`cost_model.depth_walk_applied`	`False`
`cost_model.fee_assumption.account_specific`	`False`
`cost_model.fee_assumption.assumption_basis`	`Kraken Pro Spot Crypto starter tier assumption: $0 30-day volume maker 0.25%, taker 0.40%.`
`cost_model.fee_assumption.base`	`BTC`
`cost_model.fee_assumption.fee_currency`	`quote`
`cost_model.fee_assumption.fee_model`	`maker_taker_spot_crypto`

Conformance

What this receipt supports.

It supports

A paper-only, category-level finding about the frozen Lorentzian-distance classifier on BTC/USD and ETH/USD under the stated account and cost treatment.

It does not support

A finding about a named product, vendor, author or implementation, a live-money result, or a recommendation to use the mechanism.

Checksums

The sealed export remains independently verifiable.

File	SHA-256
`xbtusd_signals.csv`	`bdaca4b3c7e5337599b527cf18e7e59c4ae775513a6aab3ce94c63b082dda590`
`xbtusd_trades.csv`	`da8ae47cec2e2028e452b57553737bcec5cbcdd772dc57ac24d395d282a25f22`
`xbtusd_equity_curve.csv`	`647084331a1d8bf24c3994f8000ddc9be935d52a3114c3473796d81fc65d0cbb`
`ethusd_signals.csv`	`f2f6a89c72621eaa538f2448bace6f3704707ffd11dd45efddde514f17a40d48`
`ethusd_trades.csv`	`b041cc3af325322cee82d42c20b67c51b29339192098f0fc03e8ba39d89a5424`
`ethusd_equity_curve.csv`	`346bd8b56b8c3cc779a8ba5d0ae3beb38b79fb757bc5387917a8f2bf22a5fe11`
`metrics.json`	`27004bbb8e351b8c76694228f646fcb369c9813b91098bd6aaf936d7d9e83975`
`receipt.json`	`702389d993be46f976ff5f1314f8bd57cd9beb3b02c84cadf1751e8d22a86a39`
`README.md`	`20bbf437dcc877190c43d5b02bc28e41591ad68f418b579da8bd936f140ae45a`
`SHA256SUMS`	`488a9ea0023231f4f7b6810985aa296b4c87484cd2aa635290b8c2753c7baf8b`
`assay-002-lorentzian-conformance.md`	`2d58153c03bbb7ba4b733664294df97af7adedbc856a418376fa93fbb2d374d0`

The source capsule's original SHA256SUMS is preserved. EXPORT_SHA256SUMS binds the Assay-side export, including the carried conformance record.

Review log

Publication record.

Date	Review	Result
2026-07-06	Gate 0 pre-registration	Frozen before evaluation
2026-07-06	Specification conformance	Pre-result conformance review passed
2026-07-08	Keel evaluation	One-shot run at engine commit `50df0c8`; both pairs falsified
2026-07-08	Assay export and publication	Sealed capsule copied one-way into Assay and published after owner review

Assay 002 evidence receipt.