Skip to content
bluebell

Strategies

Four systems, each with one job.

A strategy that tries to trade every market ends up trading none of them well. Each of these looks for one specific condition and stands aside the rest of the time.

The roster

What each one looks for.

Premise, description, what it trades, and the full record — with backtested and live phases kept visibly apart.

Strategy names, codes and descriptions on this page are placeholders. The real roster has not been supplied yet, and these must be replaced with it before the site goes live.

Example data. Every figure in this section is invented to demonstrate the product. It is not a bluebell result and not a live market reading. Modelled as at 29 Aug 2026.

BB-MOM-04as at 29 Aug 2026

Multi-horizon momentum

Buys strength that agrees with itself across three timeframes.

Looks for the fast, medium and slow exponential moving averages to align in the same direction, then waits for a range compression to resolve on expanding volume. It stands aside entirely when the timeframes disagree, which is most of the time.

Trades
Equity, Index
Horizon
20 bars
Cadence
A few signals a month
Backtested — hypothetical, benefits from hindsightLive from 6 Sept 2025
Period return
+58.11%
Max drawdown
14.10%
Sharpe ratio
1.40
Win rate
58%
Avg win / loss
2.1
Signals
214

Return and drawdown cover the whole window shown, backtest included. See methodology for how each figure is computed.

BB-TRD-02as at 29 Aug 2026

Index trend follower

Holds the primary trend in index instruments and sizes to volatility.

A slower system that takes a weekly trend filter as its permission and uses daily pullbacks for entry. Position size scales inversely with realised volatility, so a calm market gets a larger position than a violent one.

Trades
Index
Horizon
40 bars
Cadence
A handful a year
Backtested — hypothetical, benefits from hindsightLive from 6 Sept 2025
Period return
-6.98%
Max drawdown
13.06%
Sharpe ratio
-0.29
Win rate
62%
Avg win / loss
1.7
Signals
47

Return and drawdown cover the whole window shown, backtest included. See methodology for how each figure is computed.

BB-MR-01as at 29 Aug 2026

Mean reversion

Fades statistically extreme moves that arrive on thinning volume.

Measures how far price has stretched from its thirty-bar mean in standard deviations, and takes the other side when the extension is not confirmed by volume. Deliberately short-horizon, and the first to be switched off in a strong trend.

Trades
Fx, Commodity
Horizon
10 bars
Cadence
Weekly
Backtested — hypothetical, benefits from hindsightLive from 6 Sept 2025
Period return
+10.53%
Max drawdown
8.87%
Sharpe ratio
0.50
Win rate
64%
Avg win / loss
0.9
Signals
396

Return and drawdown cover the whole window shown, backtest included. See methodology for how each figure is computed.

BB-BRK-03as at 29 Aug 2026

Volatility breakout

Trades compression resolving into expansion, in either direction.

Waits for realised range to contract below its own baseline, then takes the break of that range on a volume expansion. Direction-agnostic: it is as willing to go short as long, and requires the breakout level to hold on retest.

Trades
Commodity, Equity, Fx
Horizon
20 bars
Cadence
Two or three a month
Backtested — hypothetical, benefits from hindsightLive from 6 Sept 2025
Period return
+55.92%
Max drawdown
11.62%
Sharpe ratio
1.19
Win rate
49%
Avg win / loss
2.8
Signals
158

Return and drawdown cover the whole window shown, backtest included. See methodology for how each figure is computed.

Retirement

What happens when one stops working.

Strategies decay. A system tuned to one regime will eventually meet a market it was not built for.

When a strategy stops performing we switch it off — and the record stays published, marked retired, with its full history intact. Removing it would make every remaining record look better than the truth, which is exactly the survivorship bias that makes most published track records worthless.

Retiring a system is a normal event in this business. Quietly deleting it is not.