Auto-follow/AI CTREND
MOMENTUM CORE + REGIME-DRIVEN BTC OVERLAY

HIFI TRADE AI CTREND

CTREND combines two independent layers: a relative-momentum portfolio at 0.50x gross exposure and a separate BTC position of up to 0.50x. The BTC overlay activates only in a positive medium-term regime.

CAGR
59.6%
SHARPE
1.741
MAX DRAWDOWN
−27.6%
TEST PERIOD
2020—2026
MECHANICS

How the algorithm works

01

Point-in-time universe

Each step uses the top 50 USDT contracts by the turnover available at that time, without future universe knowledge.

02

CTREND momentum

The systematic core allocates up to 0.50x gross across strong and weak instruments while limiting total risk.

03

BTC regime overlay

Up to 0.50x BTC is added when the 4H close is above its roughly 30-day-ago level; the layer is removed as the regime weakens.

04

Execution

Signals are fixed on a closed 4H candle and execution is modeled at the next hourly open.

Momentum core
up to 0.50x
BTC overlay
up to 0.50x
Average gross exposure
0.68x
Median holding time
96 hours
BACKTEST

Historical results

The strategy and BTC Buy & Hold start from the same normalized point.

EQUITY · LOG SCALE / DRAWDOWN
HIFI TRADE AI CTREND equity and drawdown versus BTC Buy and Hold
Base case: 7 bp all-in one-way costs, historical funding, a point-in-time universe and execution after a closed signal.

Strategy versus BTC Buy & Hold

MetricAI CTRENDBTC B&H
Final equity21.63x8.73x
Total return+2063.3%+773.3%
CAGR59.6%39.0%
Annual volatility29.3%60.6%
Sharpe1.7410.853
Sortino2.9081.242
Maximum drawdown−27.6%−76.6%
Calmar2.1610.509

Year-by-year results

YearAI CTRENDBTC B&H
2020+146.7%+301.7%
2021+81.5%+57.6%
2022−14.6%−65.3%
2023+116.8%+154.5%
2024+55.1%+111.8%
2025+28.1%−7.3%
2026*+28.3%−29.2%
* partial calendar year
COST SENSITIVITY

Execution scenarios

5 bp
FINAL
23.52x
CAGR
61.6%
SHARPE
1.783
MAXDD
−27.5%
7 bp
FINAL
21.63x
CAGR
59.6%
SHARPE
1.741
MAXDD
−27.6%
10 bp
FINAL
17.67x
CAGR
54.7%
SHARPE
1.635
MAXDD
−29.8%
TEST

What is included

  • Sources: Binance Vision futures/spot monthly archives; USD-M perpetuals test model.
  • Period: 01 Jan 2020—31 Jul 2026; warm-up begins 01 Nov 2019.
  • The base version uses 7 bp all-in one-way cost, historical funding and a point-in-time top 50 by turnover.
  • The test contains 4,139 trades; average holding time is 164.8 hours and median holding time is 96 hours.
RISK

Limitations

  • A backtest cannot reproduce every delay, liquidity gap or individual copy setting.
  • High historical returns do not imply a comparable future result.
  • The BTC overlay reduces exposure by a regime rule, but cannot guarantee an exit before a sharp move.
BYBIT

Connect to the algorithm

Before connecting, review the copy settings and the level of risk you can accept.

Connect on Bybit
NEXT ALGORITHMAI Momentum v.1