Strategies

This run ported the hyperopted Freqtrade 'Supertrend' strategy line for line to LINK spot on 1h bars, using six incremental Supertrend state machines plus the original ROI, stop and trailing exits. It took 894 trades from 2019 to 2026 and lost about 0.24% per trade after fees, roughly -0.04% before them, so fees were not the root cause. At backtest review, after 2 iterations, the analyst abandoned it for negative expectancy, an inverted payoff (capped wins against losses running to the deep stop) and no forecasting information in the signal; it never reached optimization. The rejection applies only to this LINK spot port, not to the original on its hyperopted pairs. A separate, unverified concern about impact-cost accounting behind the -100%/liquidation headline was passed to the operator for a ledger review.

This strategy ported freqtrade's EMA20/OBV TrendFollowingStrategy unchanged to 5-minute LINK futures, long-only, with all parameters frozen. Its full-period backtest looked viable on cost: +6.25% over 513 trades and 0.335% per trade, which cleared the fee floor. Behind that, Sharpe was only 0.262 and it was 0.95 correlated with LINK buy-and-hold. Most profit came in 2020, and out-of-sample and holdout Sharpe were 0.02 and -0.34. After 2 iterations the analyst abandoned it on a hard positive-years gate failure (4 of 7), with no parameters left to tune and paper trading not recommended.

The Research Lead adapted Myquant's SkyPark first-minute range breakout to DOT perpetuals with UTC daily sessions, keeping the source's literal flag comparisons. After 5 iterations, including a fix that fell back to bar closes when tick data was missing so the full history was evaluated, the backtest produced 2,036 short-only trades. They returned -97.3% with Sharpe -1.83, losing even while DOT fell heavily. The analyst abandoned it at backtest review for negative expectancy, noting that the verdict rejects only the DOT/UTC adaptation, not the original strategy on session-based futures.

AtomSpotTradingViewRsi translated the MIT-licensed Jesse TradingView_RSI strategy with its source defaults frozen, on ATOMUSDT spot 4H bars. The platform chose the instrument and interval because the source named neither. On a reliable sample of 727 trades it lost money in calm, normal and stressed volatility regimes alike, with signal IC of 0.000, and was wiped out in March 2021. The analyst abandoned it at backtest review, before optimization, so later stages were never reached. The verdict covers only this ATOM 4H operationalization, not the Jesse strategy in general.

This was the second iteration of a DOT/OKX daily SMA 5/15 crossing strategy, re-registered with half the exposure (10% of equity) to meet a 15% drawdown contract. The backtest met that contract (11.7% drawdown) with +8.3% over 67 long-only trades. The analyst still found no edge: Sharpe 0.24, negative alpha and IR, a signal IC indistinguishable from noise, weak out-of-sample results, and nearly all profit coming from 2021 and one month in November 2024. It failed the hard gate on positive calendar years (3 of 6) and was abandoned without paper observation.

BtcOkxFmzSuperTrend453 ported the source's fixed 45/3 SuperTrend flip mechanism to BTC OKX perpetual trading with market execution and equity-relative sizing. Trade records supported the intended alternating close-before-reopen behavior, but 5,431 closed trades produced negative net expectancy and a reported -100% return. Backtest review abandoned the strategy before optimization because losses were already substantial before reported liquidation and no verified repair supported iteration. Reporting inconsistencies were left for separate investigation.

After three iterations, the strategy tested a single-position, trend-gated pullback approach on ETH and ADA. Its 432 closed trades produced a -42.26% return and near-zero Sharpe, as ADA's losses overwhelmed ETH's positive contribution. Backtest review abandoned this adaptation for negative expectancy and insufficient per-trade return before optimization or later reviews.

The Research Lead adapted a model-free KO-PEP spread-excursion strategy from Ananova, Cont and Xu to 1-hour USEQ bars, because the paper's second-by-second NYSE midprices were not available. After 2 iterations the code matched the hypothesis, the accounting reconciled and risk stayed small (4.93% max drawdown), but over about 2 years it lost 3.25% on 138 legs with Sharpe -0.65 and an average leg return of -0.194%. The backtest review abandoned it before optimization, finding that excursions mostly ran into the time stop instead of converging and that the signal carried no information (IC -0.098). The rejection covers only the 1-hour adaptation, not the paper's original tick-level strategy.

LinkOkxSqueezeExpansionBreakout tested compression-gated long and short breakouts with ATR-based risk management in one iteration. Its 280 closed trades produced -8.31% return, Sharpe -0.085 and negative average trade returns despite 25.18% exposure matching the hypothesis's projected range. Backtest review abandoned the implementation for negative net expectancy, so optimization and later reviews were not reached.

After one iteration, DotOkxHowtraderDoubleMa tested a frozen 10/20 SMA crossover adaptation on DOT/OKX with persistent limit orders and reversal exits. The backtest reported -100% return and Sharpe -7.34, while 985 inspected pre-liquidation trades independently demonstrated severe losses. Backtest review abandoned the adaptation for negative expectancy, with no authorized parameter search available. Later stages were not reached, and execution, reporting and dependency questions remained unresolved.

After two iterations, AtomSpotJesseDonchian tested the Jesse community’s fixed Donchian rules on ATOM spot 4-hour bars. Across 113 closed trades, it lost 69.64% with a Sharpe of 0.132 and negative dollar expectancy, despite clearing the average percentage trade-return fee floor. Backtest review abandoned this instrument/timeframe translation before optimization, without drawing conclusions about the source strategy on other markets.

After one iteration, LinkFrozenTDSequential tested a frozen, long-only TD Sequential adaptation on LINKUSDT hourly spot bars. Its 577 closed trades produced a -33.88% return, -0.0581 Sharpe ratio, and 73.32% maximum drawdown. Backtest review abandoned it for negative expectancy after costs, declining optimization because parameters were frozen and no verified repair supported iteration.

After three iterations, JesseIFR2AdaSpot reached backtest review as a frozen adaptation of the Jesse community IFR2 rules. Across 261 closed trades, it returned 15.82% net with Sharpe 0.272, but maximum drawdown reached 71.15%. Review abandoned the run because its 0.24923% average net trade return narrowly missed the spot floor, modeled impact was material, and no tuning was authorized. The decision concerned this ADA four-hour adaptation; later evaluation stages were not reached.

After one iteration, DotSpotJesseSimpleBollinger reached backtest review as a faithful fixed-default, long-only DOT hourly translation. Its 847 closed trades produced a reported -0.3553% net average trade return, -1.23 Sharpe and 0.771 profit factor. Review abandoned it because gross trade returns fell below spot commissions and its fixed contract offered no legitimate optimization scope. Accounting inconsistencies were flagged separately, and the rejection applies to this DOT translation rather than the original strategy on other markets.

After one iteration, EtcOkxStretchBreakoutPreference tested a directional Stretch breakout adaptation on ETCUSDT.OKX with protective stops and same-day exits. Its backtest reported 998 closed positions, -100% return and a Sharpe of -1.6318. Backtest review abandoned it for negative expectancy before optimization, while flagging execution and reporting inconsistencies for separate investigation. The rejection applies to this adaptation, not the original source strategy or session breakouts generally.

After one iteration, BchOkxDailyMacdEmaRestart tested daily MACD direction filtering with 15-minute EMA recross entries and hourly ATR risk sizing. Its backtest reported 5,077 closed positions, -100% return, and Sharpe -3.709. Backtest review abandoned it because negative expectancy was already evident before reported liquidation, so optimization and later stages were not reached. Reporting inconsistencies remain unresolved and require separate accounting review.

After two iterations, the strategy reproduced the frozen long-only MACD/EMA rules on hourly LTC spot bars. Its backtest reported 1,206 closed positions, -100% return, and Sharpe -2.1184, alongside unresolved reporting inconsistencies. Backtest review abandoned it for negative expectancy and declined optimization because no parameter search was authorized.

After three iterations, the MSFT/ORCL strategy reached backtesting with synchronized opposing legs, price-corrected hedging, and frozen coefficients. Its 11 completed baskets produced -$970.54 net PnL, a -0.971% return, and Sharpe -0.045. Actual execution occurred at session opens despite the hypothesis specifying completed closes. Backtest review abandoned this baseline adaptation before optimization because evidence was sparse and economics unfavorable, without claiming to falsify cointegration strategies generally.

After two iterations, LtcSpotIcebreakerBoundedRecovery tested bounded average-cost recovery on LTC spot using completed hourly closes. Its 1,334 closed campaigns returned -22.12% with a Sharpe of -0.317, as larger losing inventories outweighed frequent smaller recoveries. Backtest review abandoned the frozen adaptation before optimization; this outcome does not establish the performance of the original source strategy.

After two iterations, the ETH adaptation reached backtest review using hourly EMA periods of 10 and 21, although the hypothesis described 10 and 20. It lost 47.24% across 4,100 closed trades, with Sharpe -1.345 and negative returns in every reported year. Review abandoned the run for negative expectancy and losing reversal exits, so optimization and later review stages were not reached. The summary could not be saved to the specified outbox because this session has read-only filesystem access.

Across two iterations, JeSuisTmLondonBreakoutXRP tested the source’s frozen breakout rules on XRP minute bars. Its 1,784 closed trades produced a reported -100% return and Sharpe of -1.537, with material losses already present before liquidation. Backtest review abandoned the strategy before optimization because expectancy was negative and no permitted tuning path supported further iteration. Post-liquidation trading and position-metric inconsistencies remained unresolved accounting issues.

SameMonthSeasonalityEtfLS pursued cross-sectional monthly return seasonality across 18 ETFs with inverse-volatility sizing. After four iterations, QA identified an execution mismatch: the hypothesis required month-end-close fills, while the platform executes these orders at the next session's open. Negative sandbox expectancy was advisory; the execution mismatch caused abandonment before later pipeline stages.

After one iteration, MyquantBollingerBanditDotSpot tested a frozen translation of Myquant's rules on daily DOT spot bars; pinned-commit parameter provenance remained unresolved. The backtest produced 27 closed trades, a -2.539% net return, and a -0.045 Sharpe, while drawdown stayed within the declared 15% limit. Backtest review abandoned the run for negative expectancy before optimization, rejecting only this tested translation and parameter set.

EthDeribitCostAwareGammaScalping aimed to test cost-aware hourly delta-band hedging of a long ETH straddle. Across three iterations, implementation remained blocked by the rolling-options base's lifecycle behavior. All six synthetic scenarios failed during setup, so no backtest or optimization results were produced. The run was abandoned in verification, leaving the trading thesis untested.