Montcrest named to institutional digital asset manager shortlist

There is a tendency in finance to search for permanence. Investors want models that can be defined, tested and then trusted for years. Strategies are often presented as if the objective were to discover a formula that finally explains the market. We started Montcrest Capital with a very different assumption: there is no permanent formula, because the market itself is never permanent.

This is particularly obvious in digital assets. Crypto trades twenty-four hours a day, seven days a week, across every major financial session. Liquidity moves from one part of the world to another, volatility expands and contracts, participants change, and relationships between assets evolve. A strategy that performed exceptionally well in one environment can become significantly less effective when the conditions surrounding it change. To us, building a static investment process for a market like that never made much sense.

Friedrich Nietzsche wrote that “the snake which cannot cast its skin has to die.” He was speaking about the ability to abandon ideas that no longer serve us, but the principle applies remarkably well to systematic investing. Every algorithm is ultimately built on assumptions about how markets behave. It may assume that certain structures tend to mean revert, that liquidity behaves in a particular way, or that specific conditions create favorable asymmetries. Those assumptions should be tested rigorously, but they should never become sacred.

That philosophy is important because there is a difference between discipline and rigidity. We want our algorithms to be disciplined. Once a strategy is deployed, identical conditions should produce identical decisions. There should be no hesitation after a losing trade, no overconfidence after a winning streak and no decision made because something simply “feels different” today. But the research process surrounding those algorithms cannot be rigid. When enough evidence shows that the environment has changed, the model has to be questioned again.

We are not building one answer

Montcrest was never designed around a single algorithm. Our research spans different horizons, from intraday execution to structural swing strategies and longer-term Bitcoin cycle exposure. Some models operate across a broader universe of major digital assets, while others intentionally concentrate on the deepest markets.

The objective is not to accumulate strategies for the sake of complexity. It is to avoid making the entire portfolio dependent on one interpretation of how markets are supposed to behave.

Markets rarely reward one behavior indefinitely. There are environments where mean reversion dominates and others where different structural forces become more important. Short-term dislocations can create opportunities while longer-term conditions remain unattractive, or the opposite can happen. We would rather have several independent sources of return, designed around different market behaviors and time horizons, than continuously add complexity to one model in the hope that it can solve every environment.

We do not believe resilience comes from making one algorithm more complicated. It comes from combining independent sources of return that respond differently to changing market conditions.

The ambition is therefore not to create an algorithm that is always right. Such a requirement would make the entire portfolio dependent on an impossible assumption. The ambition is to build an investment architecture that does not require any single algorithm to succeed all of the time.

Automation is not the edge

This distinction matters because algorithmic investing is often treated as though automation itself were an advantage. It is not.

A poorly designed strategy can be automated just as effectively as a good one. In fact, automation can make a bad strategy more dangerous because it allows bad decisions to be repeated with perfect consistency.

The real work happens before execution.

At Montcrest, our research process is deliberately adversarial. Strategies are tested on data they have never seen, subjected to walk-forward validation, compared against randomized alternatives and examined across different market periods. We also test whether an edge survives after removing the strongest trades, because a strategy whose performance depends on a handful of extraordinary outcomes may produce an impressive backtest without representing a robust investment process.

We believe a strategy should not earn capital because it looks convincing in a backtest. It should earn capital only after surviving repeated attempts to disprove it.

That mindset is important because backtests can be extremely persuasive. With enough variables, thresholds and filters, it is possible to create historical results that appear extraordinary. The difficult part is not producing a beautiful backtest. The difficult part is determining whether the result reflects persistent market behavior or simply a coincidence discovered in historical data.

Our internal principle is deliberately skeptical: every apparent miracle remains a bug until there is sufficient evidence to believe otherwise.

The algorithm provides consistency. The research process provides skepticism.

Both are necessary.

Risk before appearance

The same philosophy shapes the way we think about performance.

One of the recurring lessons from our research is that the statistic that looks best is not necessarily the one that creates the most economic value.

During the development of one Bitcoin model, for example, widening the stop increased the historical win rate from approximately 61% to 82%. On the surface, that appeared to be a substantial improvement. In reality, expectancy fell from approximately +0.23R to +0.05R per unit of risk. The system won more frequently while extracting substantially less value from the risk it was taking.

That experiment reflects a broader principle behind Montcrest. We are not interested in optimizing the number that looks most attractive in isolation. Return matters, but so do drawdown, expectancy, capacity, execution and the amount of risk required to generate that return.

A strategy can report a spectacular CAGR and still be unsuitable for a fund if the path required to achieve it involves unacceptable drawdowns. It can report an exceptional win rate while hiding poor payoff asymmetry. It can appear highly diversified under normal conditions while becoming heavily correlated during periods of market stress.

The purpose of research is to uncover those weaknesses before capital has to discover them.

This also explains why capacity is part of our investment process rather than an afterthought. As capital grows, market depth increasingly influences what can realistically be traded. At lower AUM, a broader universe of digital assets can contribute attractive opportunities. At larger scale, the portfolio naturally places greater emphasis on the deepest markets, particularly Bitcoin, Ethereum and Solana. The theoretical percentage return can decline as a result, but the ability to deploy substantially more capital becomes more valuable.

For us, that is not a compromise. It is the difference between designing a strategy and designing a fund.

Raising the standard

Montcrest Capital will never claim to have built the perfect algorithm. Markets are too complex, too adaptive and too unforgiving for that kind of certainty.

Our ambition is different.

We want to push beyond the limitations of conventional algorithmic strategies and raise the standard of what systematic investing should demand from itself. That means going further than basic automation, higher win rates or attractive backtests. It means combining robustness, drawdown discipline, scalability, adaptability and diversification into the design of the investment process itself.

In many ways, the objective is to create a level of algorithmic discipline and scrutiny that is still rarely discussed, let alone applied consistently.

We do not view adaptability as changing an algorithm whenever a few trades go wrong. That is not adaptation; it is often just overfitting. The rules of a deployed strategy must remain measurable and disciplined. What must remain adaptable is the research architecture around them: the willingness to challenge assumptions, retire ideas that no longer survive scrutiny, introduce genuinely independent sources of return and allocate capital differently as liquidity, risk and market structure evolve.

The algorithms provide discipline. The research process provides adaptability. Risk management determines how much conviction the portfolio is allowed to express, and diversification prevents one idea from becoming the entire business.

That combination is what we are trying to build.

Built to remain relevant

We cannot know what the digital asset market will look like in five or ten years. We do not think anyone can.

What we can do is build systems that do not depend on the next decade looking like the last one.

That is the standard we are trying to set at Montcrest Capital: systematic enough to be measured, disciplined enough to be trusted, and adaptive enough to remain relevant.

Nietzsche's snake survives because it is capable of shedding what it has outgrown. We believe an investment process should be capable of doing the same, not impulsively and not because the latest trade lost, but because the evidence eventually demands it.

That, more than any individual model or backtest, is the philosophy behind Montcrest Capital.