LinkDailyVolumeConfirmedDonchianBreakoutLong

LINK Daily Volume-Confirmed Donchian Breakout Long (Pure Mechanism)

A long-only, single-instrument, single-timeframe strategy on LINKUSDT.BINANCE DAILY bars implementing the PURE analyst-validated Volume-Confirmed Donchian Breakout template. Fills the last clean instrument-timeframe gap on this template: BTC Daily, ETH Daily, SOL Daily, ADA Daily/4H, XRP 4H, BNB 4H, BTC 4H, ETH 4H, SOL 4H — all variants exist; LINK Daily Volume-Confirmed Donchian does NOT. LinkDailyMultiWeekTrendContinuationLong covers LINK Daily but uses a different mechanism (multi-week trend continuation, not Donchian channel breach). The PURE template: 20-day Donchian high breakout + volume confirmation + 50-day SMA trend gate. Three filters only — keeps trade count high for walk-forward statistical power per the analyst's recurring guidance. LINK Daily provides three specific advantages: (a) LINK has distinct narrative cycles from BTC/ETH/SOL/ADA — oracle integrations, CCIP launches, BUILD partnerships, RWA tokenization events — producing Daily breakouts decoupled from broader L1 beta; (b) LINK Daily ATR is typically 2-5% of price, comfortably above the 0.10% perp fee floor; (c) LINK Daily bars are CONFIRMED WORKING via LinkDailyMultiWeekTrendContinuationLong. Single-instrument, single-timeframe, NO secondary dependencies — directly avoids the multi-instrument audit_stale failure mode that has killed every cross-instrument proposal.

Iteration 2 fix for the 'leverage_set_but_unused' gate. The config set leverage=2.0 and position_size did read leverage, but via getattr(self.config, 'leverage', 1.0), which the static gate does not recognize as a leverage reference — so it reported the leverage as inert. Fix (smallest possible): changed the sizing line to read self.config.leverage DIRECTLY (leverage = float(self.config.leverage)), the exact pattern the gate detects, and it is genuinely consumed in notional = equity * position_fraction * leverage / price. Nothing else changed: the 3-filter Donchian/volume/SMA entry logic, the Turtle exit, the look-ahead-safe prior-window slicing, the continuous tanh fallback signal, and all imports (numpy + Decimal only) are untouched, so the earlier-passing layers do not regress. Venue remains BINANCE USD-M futures at leverage 2.0; with position_fraction=0.5 the effective notional is ~1.0x equity but the leverage is now wired into sizing (not inert), so changing leverage changes the backtest as required.

Fundamentally broken — the strategy blew up the account and has no edge, so optimization cannot help. Risk metrics are catastrophic: CAGR -100%, recovery_factor -1.0, max_drawdown 85.7%, with single-day portfolio returns of -180.65% (2022-06-12) and -195.54% (2025-08-01) and annual returns 2022 -227% / 2025 -119%. Proximate cause: the 'pure 3-filter' template has NO stop-loss — the only exit is a 10-day Donchian-low break, which on daily bars fires far too late, so with leverage 2.0 a single violent down-bar loses more than the entire equity (avg_position_pct reaches 231% as equity collapses). Underneath that, there is no edge on LINK Daily anyway: profit_factor 0.69, win_rate 30.8%, expectancy -$4,827/trade over just 26 trades, with 2023/2024/2026 flat (breakout rarely fires). The +49.8% headline total_return is 175.3% open-position UNREALIZED (end_unrealized_pct=175.3) -> realized return ~ -125%. This is not a code bug in the entry path (26 clean entries) and not optimizable: adding a stop would only convert blow-ups into many small losses without creating winners. The volume-confirmed Donchian breakout template does not generalize to LINK Daily — the daily timeframe is too sparse (26 trades, multiple dead years) to validate and the realized edge is negative. The mechanism is captured on the validated asset/timeframe arms; recommend the Research Lead not extend it to LINK Daily. If anything is salvaged, the missing protective stop-loss + leverage interaction is a risk-design flaw the developer should never ship in a long-only leveraged strategy.

Long-only pure 3-filter Volume-Confirmed Donchian Breakout on LINKUSDT.BINANCE USD-M perp, daily bars, pure OHLCV. Enters long when the close breaks above the prior 20-day Donchian high, bar volume >= 1.5x the prior 20-day average volume, and the close is above its 50-day SMA. Exits on a classic Turtle 10-day Donchian-low break. calculate_signal() returns 1.0 on the confirmed 3-filter entry and a damped tanh trend-deviation proxy otherwise (kept within +/-0.5 so it never trips entry) to keep the signal varying. position_size sizes notional = equity * position_fraction * leverage / price, genuinely consuming the configured leverage 2.0 (~1.0x equity notional at fraction 0.5).

Verification failed (leverage_set_but_unused): config sets leverage=2.0 but the strategy never references leverage in sizing. Size positions on it — notional = equity * fraction * leverage / price (read self.config.leverage) — or set leverage=1.0. As written the leverage is inert and the backtest is identical to 1x.

Clean implementation of the entry logic (26 entries signaled -> 26 submitted, 0 dropped); look-ahead-safe prior-window slicing.

Single-instrument, single-timeframe — avoids the multi-instrument stale-leg failure mode as intended.

Account blow-up: CAGR -100%, recovery_factor -1.0, max_drawdown 85.7%, with single-day returns of -180.65% and -195.54% — a long-only strategy losing >100% in one bar.

No protective stop-loss: the only exit is a 10-day Donchian-low break, which on daily bars fires far too late; with leverage 2.0 the position runs past full equity (avg_position_pct 231%).

No edge: profit_factor 0.69, win_rate 30.8%, expectancy -$4,827/trade over only 26 trades; deeply negative 2021/2022/2025.

+49.8% headline total_return is 175.3% open-position UNREALIZED -> realized ~ -125%. Multiple dead years (2023/2024/2026 flat) show the breakout rarely fires.

total_return 49.8 minus 175.3 unrealized

>100% single-bar loss -> blow-up

This strategy filled the last open slot for the volume-confirmed Donchian breakout template — a 20-day breakout with volume and 50-SMA filters — on LINK Daily. The entry logic was clean over 26 trades, but the 'pure' design carried no stop-loss, and with 2.0x leverage and only a slow 10-day Donchian-low exit, single down-bars of -180% and -195% blew the account through zero (CAGR -100%, 85.7% drawdown), while the +49.8% headline was 175% unrealized MTM masking a ~-125% realized result with no underlying edge (PF 0.69). The analyst abandoned it at the backtest-review gate on its second iteration as fundamentally broken and non-optimizable — flagging the missing stop-loss/leverage interaction as a risk-design flaw and recommending the template not be extended to LINK Daily — so it never advanced to optimization or risk review.

A leveraged long-only strategy must never ship without a protective stop — relying solely on a slow 10-day Donchian-low exit let a single daily down-bar lose >100% of equity — and a template validated elsewhere can both fail to generalize (sparse, edgeless on LINK Daily) and be structurally unsafe; a +49.8% headline that is 175% unrealized hides a ~-125% realized blow-up.

The backtest-review analyst issued an 'abandon' verdict: the entry path was clean but the 'pure' template ships NO stop-loss — its only exit is a slow 10-day Donchian-low break — so at 2.0x leverage a single violent down-bar lost more than the entire equity (avg position reached 231%); beneath the blow-up there was also no edge (PF 0.69, multiple dead years), and the template does not generalize to the sparse LINK Daily timeframe.

A long-only, single-instrument 'pure' 3-filter Volume-Confirmed Donchian Breakout on LINKUSDT daily Binance futures (2.0x leverage) — entering long on a 20-day Donchian-high breakout with volume confirmation (≥1.5x the 20-day average) above a 50-day SMA trend gate, exiting only on a 10-day Donchian-low (Turtle) break — extending a template validated on other asset/timeframe arms to the last open LINK Daily slot.

Over only 26 trades it blew up the account: CAGR -100%, recovery factor -1.0, an 85.7% max drawdown, profit factor 0.69, win rate 30.8%, expectancy -$4,827/trade, with single-day portfolio returns of -180.65% and -195.54%; the +49.8% headline total return was 175.3% unrealized open-position MTM, so the realized return was roughly -125%.