DOT Spot Jesse SimpleBollinger — Faithful Fixed-Default 1H Translation

Hypothesis

Translate SimpleBollinger by the Jesse community, whose source credits gabrielweich/jesse-strategies, under the supplied MIT licence. Trade DOTUSDT.BINANCE_SPOT on 1-hour bars, long-only at 1x. Preserve the source's entire-balance sizing, Bollinger breakout entry, Ichimoku entry filter and middle-band exit. No additional stop, target, volatility sizing, cooldown or trend-loss exit. Source: [pinned SimpleBollinger](https://github.com/jesse-ai/example-strategies/blob/7c91e0a37bf62165790120d730442e4f6eb00364/SimpleBollinger/__init__.py). Its performance claims are unverified.

The explicit faithful-translation request determines the mechanism and direction. Spot suits its unleveraged long-only execution and helps the underrepresented spot bucket, although long-only and single-symbol experiments are already overrepresented. DOT avoids the saturated BTC series; corpus_concentration is empty, so its actual research share is unknown. This is breakout continuation, rather than the failed Connors RSI reversal mechanism. Do not change its rules to manufacture novelty or rescue disappointing results.

negative_expectancy: decline optimization of this fixed-default DOT 1H translation. Across 847 closed trades, the mean gross price move is only 0.1873%, below spot round-trip commissions before modeled impact; reported net expectancy is negative and PF is below 1. This rejects this DOT translation, not the original author's strategy on other markets. Review the accounting inconsistencies separately.

Implementation

Fixed-default, long-only DOT spot translation of Jesse SimpleBollinger on hourly bars. Enters when close exceeds the 20-period hl2 upper Bollinger band and both Ichimoku cloud spans. Exits only below the middle band. Deploys the entire available balance less Jesse's fee reserve, with quantity floored to three decimals.

Verification Results

Optionally recompute the 20-bar hl2 mean/std directly from self._bars[-bb_period:] on each bar. That costs the same and removes both the gap and the drift risk.

The Bollinger window is kept as an incremental hl2 deque, updated only when calculate_signal runs. Backtests call it on every bar after min_bars. In live trading, the base template's staleness guard skips calculate_signal on lagging bars, so the deque could miss bars and compute the bands on a non-contiguous window. The rolling sum-of-squares variance also builds up floating-point drift over very long runs. At DOT prices of about 1-50 this is negligible.

No change, to keep the translation faithful. The analyst should judge drawdown against the constraint.

Each entry deploys 99.7% of the USDT balance with no stop, as the hypothesis requires (faithful port). This is a CASH spot account at 1x, so gross is capped at 1.0x equity. No short or negative size is possible, and the notional is far above the $5 minimum. Drawdown is limited only by the middle-band exit on bar close, so the pre-registered 40% max-drawdown constraint is likely to bind.

Sandbox (557 days, 318 closed long trades, unoptimized fixed defaults): avg_trade_return_pct -0.51% against the 0.20% spot round trip, PF 0.62, win rate 32%. Commissions were 21% of gross. Fee arithmetic: 318 trades x 0.2% is about 64% of notional in fees alone, and the hourly 2-sigma breakout with a middle-band exit gives up roughly half a band-width per trade. The hypothesis states no per-trade gross move that clears 0.2% with margin. The headline -100% / 'liquidated' on a 1x CASH account does not come from leverage: the strategy compounds losses at about 99.7% balance exposure, and the folded impact cost ($104k) is larger than the depleted equity. The avg_position_pct of 1247% is a metrics artifact: the equity-before denominator collapses near zero late in the run. Per-trade sizing never exceeds 0.997 x the USDT balance. Check first: per-trade gross before costs in the full backtest. If it is below about 0.2%, the analyst should abandon; the hypothesis pre-registers fixed defaults with no tunables, so optimization cannot help.

Backtest Review

Sharpe

-1.23

Total return

-100.00%

Max drawdown

100.00%

Trades

847

Win rate

29.0%

Profit factor

0.77

847 closed trades are all long, consistent with the declared direction; code implements breakout entry, cloud filter and middle-band exit.

Audited hourly inputs cover the intended evaluation window.

QA's economic concern is borne out: mean entry-to-exit price return before costs is 0.1873%, below the approximately 0.20% spot round-trip commission. Reported net average trade return is -0.3553%, versus the 0.25% spot review floor; PF is 0.771.

All contract parameters are fixed and there are no tunables. Optimization cannot legitimately rescue this translation by changing its rules.

Accounting needs separate review: 519 trades begin after reported liquidation, closed-trade net losses sum to $328,758 on a $100,000 account, and average position exposure conflicts with sub-1x peak leverage. These inconsistencies limit interpretation of the -100% headline; their cause is unverified.

The sampled buy-and-hold curve returns -84.79% with approximately 98.54% drawdown. This comparison does not establish an active edge.

Outcome Summary

This fixed-default DOT translation failed to cover trading costs, while its accounting inconsistencies require separate review.

Backtest review returned abandon because the mean gross price move of 0.1873% was below approximately 0.20% round-trip spot commissions before modeled impact. All parameters were fixed with no tunables, so optimization was declined; optimization and subsequent review stages were not reached.

Faithfully translate SimpleBollinger to long-only DOT spot trading on hourly bars at 1x, using entire-balance sizing, Bollinger breakout entries, an Ichimoku filter and middle-band exits.

Across 847 closed trades, reported net average trade return was -0.3553%, Sharpe was -1.23 and profit factor was 0.771. Reported total return and maximum drawdown were -100% and 100%, respectively, but the reviewer flagged accounting inconsistencies that limit interpretation of those headlines.

Analysis

SimpleBollinger