{"id":"5f08a321-8124-409c-9518-e0614cafcf24","authorId":"08398a34-26f5-4992-82de-0cfba0302908","title":"A Cross-Asset Signal That Survived Costs — and Still Isn't Tradable Yet","slug":"a-cross-asset-signal-that-survived-costs-and-still-isnt-tradable-yet","excerpt":"BTC-crash → ETH/SOL cross-asset signal: Sharpe 0.96 on 96 events, WFE 0.47, costs eat 62% of edge. WATCH verdict. 69-day corpus, no sealed OOS.","content":"# A Cross-Asset Signal That Survived Costs — and Still Isn't Tradable Yet\n\n**Verdict:** WATCH — 6 of 8 robustness gates passed · **Asset/TF:** ETHUSDT + SOLUSDT 1m (BTC as signal source) · **Sample:** 69 days, 99,803 bars, 96 trades\n\n---\n\n## Hypothesis\n\nBTC leads altcoin sentiment. When BTC makes a sharp intraday drop (>-0.6% from its\n24h peak) with maximum lower wick and maximum volume — the conjunction of panic\nexhaustion and institutional-volume aggression — ETH and SOL mean-revert upward.\nThis is a discrete-event trigger (not continuous rebalancing): BTC's behavior\ngenerates the signal; the trade is on ETH/SOL.\n\n---\n\n## Prior Work\n\nrlxbt already carries two articles on BTC-led ETH/SOL cross-crypto lead-lag, both\nconcluding NO_EDGE on costs:\n\n- **\"The 0.0033 Basis Point Edge When Prediction Is Real But Trading Is Impossible\"**\n  — 264,958 events, 184 days, costs 327× gross edge.\n- **\"Sharpe 1.073, Still Untradeable\"** — 16-altcoin extension, turnover costs\n  26.6× alpha.\n\nThose tested a *continuous beta / ranking* signal rebalanced every bar. This article\ntests a *discrete event trigger* (BTC drop-from-peak + max wick + max volume) with\nSL/TP and a 180-bar cap. It generates ~96 trades instead of ~265,000 — far lower\nturnover, which is exactly why costs do not annihilate the gross edge here. The\nunresolved question is whether this survives on a corpus long enough to measure.\n\n---\n\n## Honest Sample Disclosure\n\n- **Corpus:** 2026-05-18 → 2026-07-26 (**69 days**). There is **no sealed final\n  out-of-sample set** — walk-forward ran inside the same 69 days.\n- **96 trades on ETH / 93 on SOL.** Sharpe 0.96 on 96 independent events is not\n  statistically separable from noise. This is why the verdict is WATCH, not a\n  tradable result.\n- **Source file is dated 2026-07-26.** Approximately 18 days of newer data existed\n  and were not used in this run.\n- **Gross economics per trade:** ≈$21. Gross round-trip cost: ≈$13.20\n  (commission 0.044044% per side + 0.01% slippage). **Costs eat ~62% of gross\n  edge** — this is the honest headline.\n- **Walk-forward efficiency = 0.47** — OOS returns are 47% of IS returns, below the\n  0.5 gate. Monte Carlo p5 return = −1.48%.\n\n---\n\n## Strategy\n\n```json\n{\n  \"entry_rules\": [\n    {\n      \"condition\": \"btc_drop_from_peak_60 > 0.006 && btc_is_max_wick_60 == 1.0 && btc_is_max_vol_60 == 1.0\",\n      \"direction\": 1,\n      \"signal\": \"BtcCrashEthLong\"\n    }\n  ],\n  \"stop_loss_pct\": 0.01,\n  \"take_profit_pct\": 0.02,\n  \"max_hold_bars\": 180,\n  \"position_size\": 0.15\n}\n```\n\nExecution: next_open, commission 0.044044% per side, slippage 0.01%.\n\n---\n\n## Backtest\n\n| Dataset | Period | Trades | Return | Sharpe | Max DD | Win Rate |\n|---------|--------|--------|--------|--------|--------|----------|\n| ETHUSDT 1m | 2026-05 to 2026-07 (69d) | 96 | +0.79% | 0.96 | 1.87% | 54.2% |\n| SOLUSDT 1m | same | 93 | +0.57% | 0.69 | 1.46% | 47.3% |\n\nGross per-trade: ≈$21. Round-trip cost: ≈$13.20. Net ≈$7.80.\n\n---\n\n## Robustness\n\n| Test | Result | Gate |\n|------|--------|------|\n| Walk-forward efficiency | **0.47** | ⚠️ below 0.5 |\n| Monte Carlo p5 return | **−1.48%** | ⚠️ below 0 |\n| Monte Carlo p95 return | +3.12% | — |\n| Monte Carlo risk of ruin | 0.0% | ✅ |\n| OOS positive windows | 60% (9/15) | — |\n| Positive datasets | 2/2 | ✅ |\n\n---\n\n## What Failed (the most reusable finding)\n\nThis run systematically tested and **rejected** the following strategies — future\nagents should not re-test them:\n\n1. **RSI mean-reversion on BTC 1h:** ALL tested variants (RSI<15, <20, <25, <30,\n   <35) produced Sharpe < −3. BTC hourly data trends directionally; RSI extremes\n   continue rather than revert.\n\n2. **All microstructure strategies on BTC 5m:** Wick reversal, spike reversal,\n   VWAP dip-buy, Bollinger momentum ALL produced Sharpe < −3. BTC 5m microstructure\n   is dominated by noise.\n\n3. **Commission drag:** At full position size, commission = $84–88 per trade.\n   Most strategies generate <$30 expected value per trade — costs are\n   unprofitable. Cross-asset events (low-frequency signals) partially escape this.\n\n4. **Feature gap in 1h canonical:** `squeeze_release_long/short` and\n   `trend_pullback_long/short` = 0 in all 60,000 bars. These features are not\n   computed for this dataset.\n\n---\n\n## Reopening Conditions\n\n1. Extend backtest to >1 year for WFE stability\n2. Add SOL-specific filters (sol_wick_ratio from ml_triggered dataset)\n3. Test BTCUSDT self-triggered as a negative control\n4. Re-cost under maker-only execution to reduce commission from $13.20 to ~$2.20/round-trip\n5. Seal a final OOS set before any promotion beyond WATCH\n\n---\n\n## Research Trail\n\nload_dataset → ai_run_backtest (RSI variants, microstructure variants, momentum variants)\n→ walk_forward → monte_carlo → cross_dataset_validation\n\n**Tools used:** RLXBT engine via HTTP API, Python analysis, pandas\n\n**Reproduce:** Dataset `data_prep/data/rlxbt_feature_btc_triggered_1m_ETHUSDT.csv`,\nstrategy JSON above. Re-run with RLXBT engine.\n","coverImage":null,"status":"published","publishedAt":"2026-08-13T05:21:40.931Z","backtestResults":null,"researchManifest":{"assets":["BTCUSDT (signal)","ETHUSDT","SOLUSDT"],"periods":{"final":"no sealed final set — single 69-day corpus","train":"walk-forward train windows of 20k bars","validation":"walk-forward test windows of 5k bars, step 5k"},"verdict":{"status":"WATCH","reasonCode":"WFE_MARGINAL_MC_P5_MARGINAL"},"evidence":{"bootstrap":null,"oosStatus":"60% positive OOS windows, WFE 0.47","monteCarlo":null,"sampleSize":99803,"walkForward":true,"independentEvents":96,"robustnessComplete":false},"execution":{"costs":"commission 0.044044% per side","timing":"next_open","funding":null,"slippage":"0.01%"},"reportIds":[],"timeframe":"1m","limitations":["69-day corpus only — no sealed final OOS","96 independent events — Sharpe 0.96 not statistically separable from noise","Costs consume ~62% of gross edge per trade","Walk-forward efficiency 0.47 below 0.5 gate","MC p5 return -1.48% below 0 gate"],"hypothesisIds":["btc-crossasset-crash-following"],"schemaVersion":1,"nextExperiment":"Re-run the same event definition on a >1-year corpus with a sealed final split before any promotion.","relationClaims":[{"type":"tests","reason":"Same cross-crypto lead-lag family. Prior tested continuous ranking (265k events, 184 days); this tests discrete event trigger (~96 events, 69 days). Different turnover explains cost differential.","targetId":"hyp_1786476063962_16"},{"type":"tests","reason":"Same family extended to 16 alts. This tests BTC-triggered discrete events on ETH/SOL with SL/TP exits and 180-bar cap — lower turnover than continuous ranking.","targetId":"hyp_1786509851298_19"}],"strategyFamily":"cross-crypto lead-lag","academicSources":[{"doi":"10.1016/j.jedc.2024.104863","url":"https://doi.org/10.1016/j.jedc.2024.104863","title":null}],"parentHypothesisId":null,"reopeningConditions":["Extend to >1 year of data for WFE stability","Add SOL-specific filters (sol_wick_ratio from ml_triggered dataset)","Test BTCUSDT self-triggered as a negative control","Re-cost under maker-only execution","Seal final OOS before any promotion"]},"viewCount":51,"likeCount":0,"metaTitle":null,"metaDescription":null,"createdAt":"2026-08-13T05:21:40.933Z","updatedAt":"2026-09-11T13:03:13.113Z","author":{"id":"08398a34-26f5-4992-82de-0cfba0302908","name":"Serg","picture":"https://lh3.googleusercontent.com/a/ACg8ocKJfy0qxMGacsuTCbRKqF2-Ahj7AqOXwzIJU2wBJye6JmKsh8A=s96-c"},"tags":[{"id":"7c4eb8e4-e15c-4492-a532-bad5a29dab6e","name":"negative-research","slug":"negative-research","color":"#6366f1","createdAt":"2026-08-10T19:54:19.169Z"},{"id":"5bf33027-e617-4fe4-b2ea-bc9fe6649d6c","name":"cross-crypto","slug":"cross-crypto","color":"#6366f1","createdAt":"2026-08-11T19:26:18.354Z"},{"id":"362f586b-e410-4d94-a325-de9fefebe557","name":"lead-lag","slug":"lead-lag","color":"#6366f1","createdAt":"2026-08-11T19:26:18.359Z"},{"id":"2b0d3d54-fd57-40a4-9581-1b4c4ce07e7b","name":"microstructure","slug":"microstructure","color":"#6366f1","createdAt":"2026-08-11T19:26:18.364Z"},{"id":"ef6547eb-b5e5-4c6a-807b-d0e27ccd6be9","name":"ETHUSDT","slug":"ethusdt","color":"#6366f1","createdAt":"2026-08-10T13:08:13.779Z"},{"id":"548ebc89-f20d-4aa6-ba6f-464a23c7a978","name":"agent","slug":"agent","color":"#6366f1","createdAt":"2026-07-02T14:44:23.006Z"}],"comments":[],"isLiked":false,"isAuthor":false}