01Retained
Dual EMA 50/200 · retained
- Return
- +165.92%
- Sharpe
- 0.769
- MaxDD
- -29.37%
5bp funding-adjusted research accounting
BTC Final Research System V1
A frozen BTC research system now has both a historical result and a forward clock.
Daily Dual EMA 50/200 was selected from a frozen search, survived a separate deep-validation contract, and was then handed into an append-only public-data runtime. The historical result is strong enough to study and uncertain enough to keep in its lane.
Retrospective result
The research engine finally produced something falsifiable
FULL covers 1 Jan 2022 through 30 Jul 2026 under 5bp per transition side and exact authenticated funding. These are research metrics, not a live track record.
- +165.92%
- +23.83%
- 0.769
- -29.37%
02Reference
BTC price-only buy & hold
- Return
- +38.33%
- Sharpe
- 0.394
- MaxDD
- -66.94%
Exact-period reference
03Reference
BTC perpetual long · funding-adjusted
- Return
- +2.27%
- Sharpe
- 0.265
- MaxDD
- -68.12%
Exact-period reference
10bp stress
+165.12%
Sharpe 0.767 · MaxDD -29.37%
Ex-2024 return
+35.56%
Strongest full calendar year neutralized
EMA neighborhood
8 / 8 positive
6 / 8 also beat B&H Sharpe in LATER and FULL
Validation funnel
The result had to survive a sequence, not a screenshot
Search, validation, freeze, and runtime handoff were separated so later layers could not rewrite earlier evidence after seeing the outcome.
- 01
Frozen search
Three low-turnover long/flat candidates were defined before the accepted-source search ran.
- 02
Single survivor
Daily Dual EMA 50/200 was the only candidate to pass all eight frozen search gates.
- 03
Deep validation
The retained candidate passed all thirteen frozen retention gates, including 10bp cost stress and neighborhood continuity.
- 04
Strategy freeze
EMA periods, warmup, daily aggregation, state mapping, and next-5m-open timing were frozen into a retained manifest.
- 05
Forward handoff
Historical state was bridged into the public-data runtime without publishing bridge bars as forward evidence.
- 06
Prospective observation
The first retained forward research observation was recorded from the 08:30 UTC bar on 22 Aug 2026.
Prospective forward boundary
The historical system crossed into a new clock on 22 Aug 2026
The handoff preserved causal strategy state while preventing historical bridge bars from being relabeled as prospective evidence.
- 22 Aug 2026 · 08:29:51 UTC
- 6,626 completed 5m bars
- 08:30:00 UTC
- FLAT · from 08:35 UTC
Research onlyAppend-only forward ledgerCheckpoint-backed stateNo retroactive forward rows
Historical seed
30 Jul · 08:10
Immutable retained state
Bootstrap context
22 Aug · 08:20
Public completed bars only
Handoff context
22 Aug · 08:25
Not forward evidence
Forward
22 Aug · 08:30
Effective from 08:35
Frozen strategy
Simple rule, heavy evidence plumbing
The retained strategy itself is intentionally interpretable. Most of the engineering exists to make the input clock, state history, accounting, and handoff auditable.
Scientific definition
- BTCUSDT · Binance USD-M · completed 5-minute factual bars
- Exact UTC-day aggregation with 365 completed UTC-day warmup
- EMA50 and EMA200, each SMA-seeded before recursive updates
- EMA50 > EMA200 → LONG_RESEARCH_STATE; otherwise FLAT_RESEARCH_STATE
- Daily decision becomes effective only at the next exact 5-minute open
- No short, leverage, position sizing, stop, take-profit, broker, or order surface
What remains uncertain
- 01Only three completed primary trades exist in the retrospective sample.
- 02One long trend accounted for 97.4% of positive completed-trade log growth.
- 03Paired block-bootstrap Sharpe-difference intervals cross zero in both LATER and FULL partitions.
- 04The historical sample participated in research and candidate selection; this is retrospective post-selection evidence, not untouched OOS confirmation.
- 05Fixed transition friction does not include additional slippage or market impact.
- 06Forward observations are research-state records only. They are not orders, positions, entries, or a live performance track record.
Interpretation
A retained research system is still not a trading approval
The strongest public claim is narrow: one frozen BTC research system produced an attractive retrospective result and is now accumulating prospective research-state observations under an explicit forward boundary.
Historical evidence, exact accounting, explicit robustness tests, frozen state, and prospective observation are all operational.
No broker, order routing, entry permission, short permission, leverage, sizing, paper trading, or live trading exists in this system.