DotDailyPullbackToFiftyEmaUptrendLong

DOT Daily Pullback-to-50EMA in Confirmed Uptrend Long

A long-only single-instrument trend-continuation strategy on DOTUSDT perpetual futures using daily bars. Architecturally distinct from every existing pipeline strategy: rather than entering on a breakout (BNB 4H), an EMA crossover event (ETH Golden Cross), a vol regime entry (pending AVAX), a momentum breakout (pending SOL), or a deep drawdown DCA (BTC Spot), this strategy enters on shallow pullbacks to the 50-day EMA inside an already-confirmed uptrend, taking only confirmed reversal-bar signals. The mechanism: in healthy crypto majors, intermediate uptrends consistently produce sequences of pullback-and-continuation moves where buying the touch of the rising 50-EMA captures the next leg up. The combination of strict regime filtering (50-EMA already above 200-EMA AND 50-EMA itself rising for 14 days) plus a shallow-pullback entry (within 1×ATR of the 50-EMA) plus a same-day reversal-candle confirmation isolates only high-probability continuation setups. Daily timeframe, ATR trailing stop, long-only, simple parameters (~5 numbers). DOT specifically because the symbol has not been used in any existing or pending strategy, has 5+ years of clean Binance perp history, and exhibits classic L1-token cycle dynamics (sharp trend phases punctuated by shallow consolidations).

Iteration 2 targets the sole reported defect: too few trades (5 in 1,884 bars) to evaluate or optimize. Per feedback, the binding constraint was requiring the SAME bar to touch within 1×ATR of the 50-EMA AND be a close>open reversal AND close in the upper half of the range, all inside the regime. The minimal fixes: (1) decoupled the touch from the confirmation — a touch within pullback_atr of the rising 50-EMA now ARMS an entry_window (3 bars), and entry fires on a later reclaim (close back above the 50-EMA); (2) widened pullback_atr 1.0→1.5 so normal dips qualify; (3) relaxed the reversal from a three-filter AND to an OR (close>open OR upper-half close); (4) shortened rise_lookback 14→7 so fresh legs are not disqualified. The continuous ATR-normalized-distance signal, regime gate structure, ATR trailing-stop exit, and capital-relative sizing (all of which passed Layers 1-3) are untouched, and imports remain numpy+nautilus only so earlier layers stay green. Only trade-generation logic changed, raising the trade count into an optimizable range while preserving the buy-the-pullback edge.

Not worth optimizing: 8 trades over 5.8 years (8 signals / 1,925 bars) is far too small a sample — the walk-forward optimizer would fit noise. The base is also net-losing (return -53.6%, Sharpe -0.34, PF 0.76, negative expectancy/alpha). The trade scarcity is structural to the ASSET, not a tunable filter: the strategy's core confirmed-uptrend regime gate rarely holds because DOT has been in a persistent downtrend for most of its history (no trades in 2020/2022/2023/2026), and DOT was selected purely for novelty ('symbol not used') rather than for exhibiting sustained uptrends. This is already iteration 2 of the entry logic with no improvement in sample size, and loosening the gate further would defeat the 'only trade confirmed uptrends' thesis. FAILURE PATTERN: a long-only pullback-in-uptrend mechanism cannot be validated on an asset that lacks sustained uptrends; asset chosen for novelty rather than fit-to-mechanism yields a too-sparse, negative-edge sample that optimization cannot rescue. The mechanism would need a genuinely trending instrument, which is a hypothesis-level retarget, not a developer parameter iteration.

Long-only daily pullback-to-50EMA trend-continuation on DOTUSDT.BINANCE USD-M futures using OHLCV only. In a confirmed up-regime (EMA50>EMA200 and EMA50 rising over rise_lookback days), it ARMS when a bar's low dips within pullback_atr×ATR of the rising 50-EMA, then ENTERS long on a reclaim (close back above the 50-EMA) within entry_window bars, confirmed by a light OR filter (close>open OR close in the upper reversal_frac of the range). Exits via an ATR trailing stop (initial stop atr_mult_init_stop×ATR, trailed atr_mult_trail×ATR) or a max_hold_bars time stop. Capital-relative sizing capped at position_pct notional and reduced so initial-stop risk stays within risk_pct of equity.

Clean, simple, differentiated construction (pullback-to-rising-50EMA reclaim rather than a breakout/crossover entry)

Correctly long-only with 0 shorts; trades match the stated long-only direction; drawdown modest (8.2%) and exposure low (~20%)

Only 8 trades in ~5.8 years (8 signals / 1,925 bars) — far too small a sample for meaningful 3-phase optimization; the optimizer would fit noise

Negative edge on the trades taken: total_return -53.6%, Sharpe -0.34, profit_factor 0.76, expectancy -$271/trade, negative alpha

Trade scarcity is structural to DOT, not a tunable filter: the confirmed-uptrend regime gate (50>200 EMA & rising) rarely holds because DOT has been in a persistent downtrend for most of its history (no trades in 2020/2022/2023/2026)

Already at iteration 2 of the entry logic and still only 8 trades — no evidence a further loosening pass yields a meaningful, positive-edge sample without abandoning the core 'confirmed-uptrend' thesis

The entry gate is far too restrictive — 5 trades in 1,884 daily bars means we cannot evaluate the edge and optimization would just fit 5 data points. Loosen the conjunction so the signal fires often enough to assess (target ~30-50+ trades over the period): (1) Decouple the pullback touch from the reversal candle — instead of requiring the SAME bar to both touch within 1xATR of the 50-EMA AND be a reversal candle, arm on a touch within `pullback_atr` of the rising 50-EMA, then enter on a reclaim (close back above the 50-EMA) within the next N bars. (2) Widen `pullback_atr` from 1.0 to ~1.5-2.0 so normal dips qualify. (3) Relax the reversal confirmation — `reversal_frac` 0.5 plus close>open plus the touch is three filters at once; try requiring just close in the upper half of the range OR just close>open, not both. (4) Consider relaxing the 'rising for 14 days' regime to a shorter lookback (e.g. 5-7 days) so it doesn't disqualify the start of fresh legs. Re-backtest and confirm the trade count is high enough (30+) to support a meaningful optimization before proceeding.

DotDailyPullbackToFiftyEmaUptrendLong tried a clean, differentiated idea — buying shallow pullbacks to a rising 50-day EMA and its reclaim inside a confirmed DOT uptrend — with DOT picked mainly because no other pipeline strategy had used it. The backtest exposed the flaw immediately: its strict confirmed-uptrend gate fired only 8 times in ~5.8 years (no trades in 2020, 2022, 2023, or 2026) and those trades lost money, posting -53.6% return, Sharpe -0.34, and profit factor 0.76. Because the scarcity was structural to an asset that has mostly trended down — and this was already the second iteration of the entry logic with no improvement in sample size — the reviewer abandoned it at the backtest-review gate before any optimization or analyst evaluation. The takeaway recorded was that the mechanism needs a genuinely trending instrument, a hypothesis-level retarget rather than a parameter fix.

A long-only pullback-in-uptrend mechanism cannot be validated on an asset chosen for novelty rather than fit-to-mechanism; if the instrument lacks sustained uptrends, the regime gate almost never fires and yields a too-sparse, negative-edge sample that no parameter tuning can rescue.

It was abandoned at the pre-optimization backtest-review gate: with only 8 trades, a negative edge, and a profit factor of 0.76, the sample was far too small and unprofitable to optimize, so the 3-phase optimization and analyst stages were never reached. The trade scarcity was judged structural to DOT — its confirmed-uptrend regime rarely held because the asset has been in a persistent downtrend for most of its history.

A long-only daily trend-continuation strategy on DOTUSDT futures that bought shallow pullbacks to a rising 50-day EMA (on a reclaim above it with a reversal-bar confirmation) only inside a confirmed uptrend (50-EMA above 200-EMA and rising), aiming to capture the next leg up in an established trend.

The base backtest lost money on a tiny sample: total return -53.6%, Sharpe -0.34, profit factor 0.76, negative expectancy (-$271/trade) and negative alpha across just 8 trades in ~5.8 years (37.5% win rate), with only a modest 8.2% max drawdown and ~20% exposure.