Six problems, six machines: why MARS is not one insight.

The single-insight product

One idea, executed well, is the normal shape.

The overwhelming majority of trading tools are organised around one insight: a journal that makes record-keeping frictionless, an indicator that reads structure, a calculator that sizes positions. This is a good shape for a product. It is learnable in an afternoon, explainable in a sentence, and it either works for a trader or does not. MARS is deliberately not built this way, and the deviation carries real costs that are worth naming before the benefits.

What the six-machine design costs

A steeper first week and a harder sentence to say.

Designing around six failure modes produces a system with a genuine learning curve and no one-line pitch. A trader cannot adopt a quarter of it and get a quarter of the benefit, because the machinery interlocks: exposure accounting feeds the throttle, which respects the gate, which reads drawdown, which is measured against the benchmark. The costs are concentrated at the start — more to learn, more to set up, more to explain to someone else — and they are the honest reason the system is not for everyone. There is also an ongoing cost that the setup framing understates — six machines produce more readings, and more readings require the discipline of not acting on each one, which is a skill the single-tool user never has to develop.

• No single-sentence pitch, because there is no single insight.
• Partial adoption yields less than proportional benefit.
• The setup week is real work, not a download.

What it buys

The failure modes do not queue politely.

The return on that cost is coverage during the periods when several things go wrong together, which is when accounts are actually lost. A drawdown does not arrive alone — it arrives with the sizing temptation, the attribution confusion about which branch caused it, and the variance-versus-decay question, all in the same fortnight. A single-insight tool handles its one dimension excellently and is silent on the other three at precisely the moment all four are live. Six machines exist because the six problems co-occur, and co-occurrence is the condition the system was designed for.

Why six specifically

The count is descriptive, not aesthetic.

Six is not a designed number and nothing was rounded to reach it. The list is what remained after failure modes that were versions of each other were merged and modes with no distinct machinery were dropped. Each survivor earns its place by requiring a mechanism the others do not supply: exposure accounting does not detect drift, drift detection does not size positions, branch attribution does not classify regime. If a seventh distinct mode with distinct machinery were identified, the correct response would be a seventh machine rather than a defence of the number.

The interlock

The machines are more useful to each other than alone.

The design's real leverage is in the connections rather than the components. Gate state is more useful because exposure accounting keeps it honest about what is committed; drift detection is more useful because branch attribution tells it where to look; the benchmark is more useful because expectancy supplies its inputs. Any one of these extracted and sold separately would be a reasonable tool and a substantially weaker one. That interlock is also the reason partial adoption disappoints — the parts are worth less in isolation than the assembled system, which is an awkward property for a product and a correct one for a system.

Who this is wrong for

Six machines is the wrong answer to one problem.

A trader with a single, identified, isolated gap should buy the tool that closes it. Someone who knows their sizing is fine, their records are complete, and their edge is validated, but who has never priced friction, needs a friction calculation — not a governance system. The six-machine design earns its cost only when the problems are plural, entangled, and not yet fully identified, which is the ordinary condition of a trader who has been at it for a few years and cannot say precisely where the money went. The same test applies in reverse and is worth running honestly: a trader who cannot name which of the six problems they have is describing exactly the condition the six-machine design was built for, and a trader who can name one should probably go and fix that one.