Why algorithms instead of humans?

The best traders are still human. And humans have limitations.

Financial markets reward discipline, consistency and the ability to make the same decision under the same conditions.

Humans are not naturally built for that.

Even highly experienced traders experience fear, overconfidence, fatigue, hesitation and frustration. After several losses, the next valid opportunity can suddenly feel more difficult to take. After several wins, risk can begin to feel easier to justify.

The market has not changed.

The trader has.

At Montcrest Capital, this is one of the reasons we build systematic investment strategies.

But there is an important distinction:

Using an algorithm is not the edge. Building the right algorithm is.

Humans need sleep. Crypto does not.

Crypto markets operate 24 hours a day, seven days a week.

There is no closing bell.

Bitcoin can move significantly while Europe sleeps. A setup can appear during the Asian session, develop through London and resolve during New York.

No human being can monitor that market continuously while maintaining exactly the same level of concentration, reaction speed and discipline.

Sleep alone makes that impossible.

And even during waking hours, human attention fluctuates.

An algorithm does not experience that cycle.

At 3:00 AM on Sunday, it applies exactly the same rules it would apply at 2:00 PM on Tuesday.

It does not become tired.

It does not hesitate because the previous trade lost.

It does not become aggressive because the previous five trades won.

If the conditions are the same, the decision is the same.

For us, that consistency is fundamental.

Humans discover. Algorithms execute.

This does not mean human judgement has no place at Montcrest.

Quite the opposite.

Humans are exceptional at discovering ideas.

They can recognize recurring structures, develop hypotheses, understand market context and challenge assumptions.

One of our own strategies began by studying a discretionary methodology based on multi-timeframe supply and demand structures and translating it into systematic rules.

The goal was not simply to automate trading.

It was to transform an investment idea into something that could be defined, tested and reproduced.

A discretionary trader can say:

“This setup looks strong.”

An algorithm needs considerably more precision.

What defines strong?

Where is the entry?

What invalidates the thesis?

How much capital should be risked?

When should the position be reduced?

When should the trade not exist at all?

Once those questions have objective answers, the strategy becomes measurable.

But not every algorithm is a good algorithm

This distinction is frequently overlooked.

Automation does not automatically create good risk management.

An algorithm can be perfectly disciplined and still execute a terrible strategy.

Imagine a system that risks 10 units to make 1.

It may win extremely often.

Its equity curve may even look remarkably smooth for long periods.

Until one loss removes the profits from many previous winners.

The algorithm executed perfectly.

The economics were simply bad.

We observed the same principle during our own Bitcoin research.

When we widened stops, historical win rate increased from approximately 61% to 82%.

That sounds like an improvement.

It was not.

Expectancy declined from approximately +0.23R to +0.05R per unit of risk.

The strategy was winning more often while producing considerably less value for the risk it was taking.

That is precisely why Montcrest does not optimize algorithms around attractive headline statistics.

We care about how efficiently risk is converted into return.

Return is only half the equation

A strategy generating extraordinary returns with catastrophic drawdowns is not necessarily an extraordinary strategy.

For an investment fund, the path matters.

This is why metrics such as the Calmar ratio, which relates return to maximum drawdown, are important to our research.

The question is not simply:

How much did the strategy make?

It is:

How much did it have to lose along the way to make it?

One of our low-frequency Bitcoin models, for example, produced an out-of-sample research profile of approximately 78% to 83% win rate, +0.81R to +1.05R per trade, around 6% drawdown, and a Calmar ratio around 2, depending on the selection threshold.

Those numbers remain research results, not live performance.

But the objective they illustrate is central to Montcrest:

we are not trying to maximize return independently of risk.

We are trying to maximize the quality of the return.

An algorithm should survive being challenged

Another advantage of systematic investing is that every decision can be questioned quantitatively.

At Montcrest, strategies are subjected to tests specifically designed to expose false edges.

We test unseen data.

We use walk-forward validation.

We compare strategies against randomized alternatives.

We examine whether an edge survives across different periods.

We remove the best trades and ask whether the strategy still works.

And we consider whether the strategy remains deployable as capital increases.

The purpose of this process is not to make a backtest look impressive.

It is almost the opposite.

We try to break it.

If an algorithm only works because of one exceptional year, a handful of trades or a carefully optimized parameter, we do not consider that sufficient.

One algorithm is not enough either

Systematic investing also allows us to move beyond the limitations of a single strategy.

Montcrest is being built as a portfolio of independent books operating across different horizons.

Intraday.

Swing.

Long-term cycle.

Different models respond to different market structures at different moments.

This matters because the objective is not to build one algorithm that is always right.

Such an algorithm does not exist.

The objective is to combine multiple sources of expected return while controlling how much risk the portfolio takes to obtain them.

That is a very different philosophy.

The machine is not the advantage

Anyone can automate a strategy.

Automation itself is becoming increasingly accessible.

What matters is what the machine has been instructed to do.

A bad strategy automated perfectly remains a bad strategy.

A fragile backtest deployed automatically remains fragile.

A system that wins frequently by accepting disproportionate losses remains economically unattractive.

The value lies in the research underneath it:

the rules,

the risk architecture,

the validation,

the diversification,

the execution,

and the discipline to reject a strategy when the numbers do not survive scrutiny.

That is where we believe systematic investing becomes genuinely powerful.

Humans build the edge. Algorithms preserve it.

At Montcrest, we do not see algorithms as a replacement for human intelligence.

We see them as a way of protecting an investment process from human limitations.

Humans research.

Humans question.

Humans design.

Humans decide how capital should be allocated.

But once the rules are established, the machine can execute them without fear, fatigue or inconsistency.

Every hour.

Every day.

Across multiple markets simultaneously.

Humans build the edge.

Algorithms preserve the discipline required to capture it.

And in a market that never sleeps, that distinction matters.