Each year we publish an honest accounting of what worked and what did not across our programmes. This is the 2025 version, written for clients and prospective clients who want the unvarnished version rather than the annual-report version.

What worked

Q1 digital-asset positioning. The systematic regime filter inside Quant Crypto Alpha correctly identified the late-January momentum surge and the strategy added gross exposure into it. This was the single largest contributor to the year's programme-level return.

Crypto Core programme in Q2 and Q3. The long-biased core book benefited from a stable trend regime through the middle of the year, with the volatility-targeted rebalancing capturing more of the upside than a static buy-and-hold position would have. Drawdown inside the sleeve was minimal. We do not expect a repeat in 2026 because the regime is already showing signs of compression, but the quarters were what a disciplined long-biased core is supposed to look like in a benign environment.

Reduction in correlation with traditional sixty-forty portfolios. Across the full year our strategies showed a correlation of negative 0.04 to a US sixty-forty benchmark. That is our most diversifying year on record and validates the structural case we make to multi-asset allocators.

What did not

The momentum signal underperformed in Q4. A central-bank surprise in early November shifted the digital-asset regime more abruptly than the signal could adapt to. The signal correctly suppressed itself within ten sessions, limiting damage, but the Q4 contribution to the digital-asset book was negative. We have rewritten the regime-detection layer to react faster to policy-driven shifts.

The Balanced Allocation programme's cash sleeve was too small in October. This left the programme more exposed to the October digital-asset selloff than we wanted it to be. The programme still met its annual objective, but the path was uglier than it needed to be. We have widened the cash band and tightened the rebalancing trigger so that the programme can react faster to vol-regime shifts at the asset level.

Onboarding wait times for new clients increased to nine business days from a target of five. This was an operations failure, not a strategy issue, and we have added two staff to the investor relations team to fix it. We owe the clients who experienced the longer onboarding the acknowledgement that it was not acceptable.

What we changed

Retired one digital-asset signal that had been producing positive in-sample results but failed to generalise in 2024 and 2025 live performance. It had contributed negatively net of fees over the trailing twelve months. Removing it improved Quant Crypto Alpha's realised risk-adjusted return.

Added a regime-aware overlay to the digital-asset carry book to reduce exposure when the realised cross-asset correlation indicator we wrote about elsewhere on this site moves above its threshold.

Rewrote the Balanced Allocation rebalancing engine to use weekly rather than monthly triggers. Monthly was simpler operationally but produced unacceptable drift in fast-moving months.

Looking into 2026

We expect higher realised volatility across both asset classes than we saw in 2024 and 2025. Position sizing across all three programmes will be more conservative on average.

We are expanding the eligible digital-asset universe inside Crypto Core to a small number of additional large-cap tokens that have built sufficient liquidity and venue coverage over the last twelve months.

We will publish the regime state monthly so clients can see what the system is doing before it shows up in performance attribution. Transparency on the inputs is, in our view, the right complement to transparency on the outputs.

We owe our clients clear, honest accounting. This is ours for 2025.

Questions about anything in this review are welcome, on the phone or by email. We would rather discuss the rough quarters openly than gloss over them in a newsletter.