Trend Following

Strategy character

We publish the strategy's character, not its triggers. The exact entry/exit conditions are core know-how and stay private — an edge that can be trivially copied disappears fast.

Risk management: what we care about most

Markets give and take, but surviving is the only thing that lets you be there for the next big move. So risk control is not a feature — it is the foundation.

• Fixed per-trade risk cap — the worst case of any single trade is locked before entry.
• Volatility-scaled sizing — the wilder the market, the more conservative the exposure, not the reverse.
• Hard stops, no averaging down — losing positions are never "averaged" or bet on to come back.
• Multiple circuit breakers — exposure is cut automatically in extreme conditions.

Backtest Advanced

We deliberately do not advertise "compounding multiples" — such numbers ignore capacity and slippage and are easily distorted. Judge this strategy by CAGR, the annual and monthly return tables, and the equity-growth curve.

Copy-trading setup: Standard (public) vs Advanced (private upgrade)

	Standard Public copy-trading	Advanced Private upgrade
CAGR ? CAGR (Compound Annual Growth Rate) is the steady yearly return that would grow the starting balance into the ending balance over the same period. It puts strategies of different lengths on one comparable scale — a truer read of yearly earning speed than a headline "total-return multiple." Compound annual growth rate	+157%	+248%
Sharpe / Max DD	1.70 / −39%	1.73 / −49%
Win rate / Profit factor	35.3% / 1.38	35.4% / 1.33
Trade frequency	4.7/wk · 2.3-day avg hold	same
Max consecutive losses	21	22
Positive months / worst month	61% / −20.4% (Mar 2025)	62% / −26.7% (Mar 2025)
Where to copy		Invite-code registration, or a copy balance ≥ 5,000 USDT → get the copy code from support Telegram @quant_nova · QQ 3454521921

Live copy-trading accounts

Real-time equity and cumulative return of our own-capital live accounts (updated hourly). For research reference only — past performance is not indicative of future results, and this is not investment advice.

Strategy growth vs BTC buy & hold

$10,000 start · backtested compounding

Annual returns vs BTC buy & hold

2017 covers Aug–Dec; 2026 covers Jan–Jul. Advanced yearly figures are the existing backtest breakdown; see the copy-trading setup card above for this tier's headline CAGR.

Monthly returns heatmap

Each cell is that month's return — green positive, red negative, deeper = larger. Defaults to Advanced; toggle to Standard. Roughly 61–62% of months are positive. 2017 starts in August; 2026 runs through July.

How we guard against overfitting

Most miracle backtests die of overfitting. Here is our validation process and the evidence, in the open:

1. 1 In-sample / out-of-sample iron rule: every parameter was decided on data up to 2023-12-31 only; 2024 onward is validation, never tuning — any change that failed validation was discarded.
2. 2 Parameter plateaus, not lonely peaks: neighboring settings give nearly identical Sharpe, and the deployed values are deliberately not the in-sample maximum. An isolated peak is the fingerprint of overfitting; a plateau is structural edge.
3. 3 Independent matching-engine cross-check: reproduced on an independent open-source engine (NautilusTrader) at 1-minute resolution — Sharpe differs by just 0.02, trade counts match one-for-one.
4. 4 Conservative fills: stops that gap through their trigger fill at market, gaps fill at the open, real historical funding is charged — no backtest-only advantages.
5. 5 Zero-hindsight coin selection: a mechanical rule that re-picks the universe quarterly using only past data still captured 84% of the deployed pool's out-of-sample Sharpe — performance comes from the strategy, not cherry-picked lucky coins.

Out-of-sample validation (parameters frozen)

Every parameter was set on pre-2024 data; from 2024 on we only validate, never tune. The decay is fully visible — and still strong.

Questions you should ask

Who this is for

Honestly: this strategy has stretches of consecutive losses (ranges are especially painful) and it has drawdowns. Its value is not "never losing" — it is standing on the right side of the big moves, with discipline, over the long run.

The full picture: growth, drawdown & out-of-sample

Everything in one view — equity growth (vs buy-&-hold BTC) above the underwater drawdown, split at the 2024-01-01 parameter freeze into in-sample and out-of-sample. Toggle Standard / Advanced.

Our deepest drawdown, dissected

Here is our most painful stretch laid bare: 18 Jul – 7 Sep 2025, the Standard portfolio equity fell −24.0% peak-to-trough, ending the window −16.3%. BTC merely chopped sideways with false breakouts the whole time — the true risk profile: losses come from trendless, choppy markets grinding the stops, not from crashes. Conversely, the crash years (2018, 2022) were among our best. Understand and accept this out-of-sync loss rhythm before copying.

Deepest drawdown window under the microscope: 18 Jul – 7 Sep 2025

The solid line is the real BTC price (left axis); the dashed line is the whole-book strategy equity (right axis, window start = 100). BTC's own 5 trades added only −4.9% (Standard), yet the equity still fell −16.3% — the rest came from other coins chopping the same way. The markers sit on the BTC price: two false breakouts long near $122k, stopped within a day; flipped short into the drop, swept by a 90-minute bounce; the last two shorts left by the 4-bar reverse channel.

The markers are BTC's 5 actual fills that window, sitting on the real BTC price line: ▲/▼ long/short entries, red dots stop-losses, green dots channel exits. The dashed line is the whole 13-coin book's equity — not BTC alone; BTC held no position before 11 Aug, so the earlier equity drop is entirely other coins. Both tiers share the same entry/exit points.