DotSpotMultiDayDrawdownAccumulationLong

DOT Spot Multi-Day Drawdown Accumulation Long (BINANCE_SPOT, Daily, OHLCV-Only)

A long-only, single-instrument, single-venue, OHLCV-ONLY strategy on DOTUSDT.BINANCE_SPOT DAILY bars. Same proven drawdown-accumulation mechanism as the 5 landed siblings (BTC/XRP/LTC/BCH/ATOM spot DCA): buy when DOT has dropped meaningfully from its 30-day high AND is still above its 100-day SMA; sell when DOT recovers to within 5% of the peak, hits a stop, or times out. DOT fills a unique cycle profile: Polkadot ecosystem with parachain auction narrative (2021-2022 wave), substrate-based app-chains, and recent OpenGov v2 + Polkadot 2.0 upgrades. Volatility ~85% annualized. DOTUSDT.BINANCE_SPOT has been listed since 2020 — ~5 years of daily bars. Five parameters total. Hits BINANCE_SPOT venue quota gap (currently 8.8% vs ≥15% target).

Direct replication of the five landed spot drawdown-accumulation siblings (BTC/XRP/LTC/BCH/ATOM), applied to DOT to fill the BINANCE_SPOT venue quota gap. The mechanism is identical and verified: contrarian accumulation buys meaningful pullbacks (>=20% off the 30-day high) but only while the long-term regime is intact (close above the 100-day SMA), which avoids catching falling knives in structural downtrends; exits harvest the mean-reversion back toward the peak (recovery within 5%) while bounding risk with a -18% stop and a 180-bar timeout. DOT's ~85% annualized volatility makes the -20% default drawdown threshold appropriate. on_bar drives entry/exit with the richer per-trade state (drawdown vs peak, SMA regime, bar counter, position PnL) the mechanism needs, while calculate_signal returns a continuous tanh-scaled drawdown-depth score so the signal varies bar to bar. Routed to BINANCE_SPOT (not futures) because the strategy is long-only, uses no leverage, and has no funding-based edge.

Pre-optimization abandon: DOT spot drawdown-accumulation fires only 11 times over 5.8 years — too sparse to optimize (5 params fit to ~11 noisy outcomes, near-empty/zero-trade walk-forward OOS windows; 2022 already has zero trades). Headline Sharpe 2.88 is degenerate (CI [-6.6, +21.3], 64% annualized vol), and the edge is marginal/concentrated (+7.6% total over 5.8yr, carried by 2021, 2024 -8.4%). Same sparse arm of the drawdown family already abandoned on LTC (18 trades) and DOGE (15 trades); mechanism is captured by 5 landed liquid siblings. Code is correct (custom on_bar path; entry_diag 0 is cosmetic), so not a developer iterate. Not worth 2 hours.

Long-only, single-instrument, OHLCV-only drawdown-accumulation strategy on DOTUSDT.BINANCE_SPOT daily bars. Buys when DOT has dropped at least 20% from its 30-day high while still trading above its 100-day SMA (uptrend regime intact), and exits when price recovers to within 5% of the peak, a -18% per-position stop-loss triggers, or a 180-day max-hold timeout fires. Positions are sized at 20% of account equity per leg on a CASH (spot, leverage=1) account, so the strategy is structurally long-only. Five tunable parameters with hard bounds.

Positive in-sample expectancy (+$691/trade, profit_factor 1.44, 73% win rate) and correct mechanism implementation (custom on_bar path; entry_diag 0 is cosmetic, 11 trades did execute); proven mechanism on 5 liquid siblings.

Structurally too sparse: only 11 trades over 5.8 years — the -20% drawdown + 100-SMA gate fires rarely on DOT, far too few for meaningful 5-parameter optimization.

Degenerate headline metrics: Sharpe 2.88 with CI [-6.60, +21.34], annualized_volatility 64%, probabilistic_sharpe 0.71 — the Sharpe is noise.

Marginal, concentrated edge: total return only +7.6% over 5.8 years (CAGR 1.8%), carried by 2021 (+12.9%), with 2024 -8.4% and zero trades in all of 2022 (below 100-SMA); avg_loss > avg_win.

Same sparse arm of the drawdown-accumulation family already abandoned on LTC (18 trades) and DOGE (15 trades).

Not worth 2 hours of optimization. Only 11 trades over 5.8 years — structurally too sparse to optimize: fitting 5 parameters to ~11 outcomes is curve-fitting noise, and the walk-forward OOS windows would be near-empty or zero-trade (2022 already has zero trades, DOT stayed below its 100-SMA through the bear). The headline Sharpe 2.88 is degenerate (CI [-6.60, +21.34], annualized vol 64%, probabilistic_sharpe 0.71), and the underlying edge is marginal and concentrated: +7.6% total over 5.8 years (CAGR 1.8%), carried by 2021 (+12.9%) with 2024 -8.4%. This is the same structural sparsity that led to abandoning the LTC (18 trades) and DOGE (15 trades) arms of this exact drawdown-accumulation family — the mechanism is already captured by the 5 landed, more-liquid siblings (BTC/XRP/LTC/BCH/ATOM), and DOT is another sparse arm of the family's asset-selectivity. The code correctly implements the mechanism (the custom on_bar path is why entry_diag reads 0 despite 11 trades — not a bug), so this is a sparsity/premise issue, not a developer iterate. Abandon rather than spend the compute.

DotSpotMultiDayDrawdownAccumulationLong extended a proven spot drawdown-accumulation mechanism to Polkadot, citing DOT's parachain/2.0 cycle profile and ~5 years of history. The code correctly implemented the mechanism (its custom on_bar path is why entry_diag cosmetically reads zero despite 11 executed trades), and the in-sample numbers looked encouraging — 73% win rate, profit factor 1.44, positive expectancy — but the edge was marginal and concentrated, just +7.6% over 5.8 years carried by 2021 with 2024 negative and 2022 producing no trades at all. With only 11 trades the headline Sharpe of 2.88 was statistical noise (CI spanning -6.6 to +21.3), too sparse to optimize five parameters or validate out-of-sample. The analyst abandoned it before optimization as another sparse arm of the same drawdown family already abandoned on LTC and DOGE, with the durable edge living on the five more-liquid landed siblings.

A positive in-sample edge is not enough when the signal is structurally rare: ~11 trades over 5.8 years cannot support five-parameter optimization or populate walk-forward/holdout windows, so a degenerate high-Sharpe-wide-CI result on a sparse arm should be abandoned rather than tuned.

It was abandoned at the pre-optimization backtest-review gate (verdict: abandon) for structural sparsity: 11 trades over 5.8 years is far too few to fit five parameters without curve-fitting noise, and walk-forward OOS windows would be near-empty (2022 already had zero trades as DOT stayed below its 100-SMA through the bear).

A long-only, spot, OHLCV-only DOTUSDT daily drawdown-accumulation strategy — the same mechanism as five landed siblings (BTC/XRP/LTC/BCH/ATOM) — buying when DOT fell ≥20% from its 30-day high while above its 100-day SMA, then exiting on recovery to within 5% of the peak, an 18% position stop, or a 180-bar time stop.

Over ~5.8 years it fired only 11 trades, with a positive in-sample profile — 73% win rate, profit factor 1.44, +$691/trade expectancy — but a marginal, concentrated overall return of +7.6% (CAGR 1.8%) carried by 2021 (+12.9%), with 2024 at -8.4% and zero trades in all of 2022. Its headline Sharpe of 2.88 was degenerate, sitting on a CI of [-6.60, +21.34] with 64% annualized volatility, and avg loss exceeded avg win.