The MARS Story · Six Problems
Six failure modes, six pieces of machinery.
Read this first
The first three
01Judgement failures share one shape: a true fact used as the wrong evidence.
Profit is real, and it is not edge. A strong feeling about a setup is real, and it is not a sizing input. A drawdown is real, and the temptation is to treat it as a temporary inconvenience rather than a change in what the account is permitted to do. In each case the trader is not being irrational so much as reasoning from the nearest available number, and the machinery works by making a better number nearer.
The four branches
02Problem five is solved by refusing to average.
Normal is the stable compounding branch. Trend Partial balances early monetization against runner potential. Trend No-Partial preserves full exposure for fat-tail continuation, at higher variance. Overflow is supplemental and has to stay controlled, because its volume and cost profile can distort blended expectancy. Reported as one number, a healthy Normal branch can hide an Overflow branch that is bleeding, and the trader will conclude their system works.
- Normal: the stable compounding branch.
- Trend Partial: early monetization balanced against runner potential.
- Trend No-Partial: full exposure preserved for the fat tail, with stricter monitoring.
- Overflow: supplemental flow that must stay controlled.
The last one is the hardest
03Variance and decay look identical from inside a losing month.
This is problem six and it is the reason the simulation and diagnostic layers exist at all. Without a modelled distribution, a losing stretch has no reference class, and the trader must decide from feel whether they are inside normal behaviour or watching an edge disappear. Both answers are catastrophic when wrong in opposite directions: changing a working system, or defending a broken one.
Why six and not one
04Fixing any one of them alone buys time rather than safety.
The six interact. Unmeasured friction feeds false confidence in the edge, false confidence feeds emotional sizing, emotional sizing deepens drawdown, and deeper drawdown makes the variance question harder to answer honestly. A trader who solves only the sizing problem still ends up in the same place more slowly. That interaction is why MARS was built as a system rather than a technique.
Reference
The six problems and what answers each
| # | Problem | The machinery |
|---|---|---|
| 1 | Traders confuse profit with edge | Branch probabilities, expectancy, blended EV, and structural drift separate outcome from edge |
| 2 | Traders deploy risk emotionally | Risk routes through gate state, drawdown level, throttle logic, open exposure, and EV condition |
| 3 | Traders ignore drawdown structure | Gate states from Growth to System Lock define capital posture; a good setup cannot override a restricted gate |
| 4 | Traders mismanage open exposure | Open risk is subtracted from authorised pool before fresh trades are sized |
| 5 | Traders cannot tell which part is working | Normal, Trend Partial, Trend No-Partial and Overflow are measured separately, never blended into one number |
| 6 | Traders cannot separate variance from decay | Structural pipelines and benchmark bands distinguish a bad stretch from a deteriorating machine |
Connected inside MARS
This module doesn't work alone.
Go deeper
Operator briefs on this territory.
Deep dive — 01
All six problems are measurement failures wearing behavioural clothes.
Why every 'psychology problem' on the list shrinks to the size of its data once instrumented.
Read the full brief →
Deep dive — 02
Confusing profit with edge is the first problem for a reason.
Why the profit-equals-edge confusion sits at the head of the list, and what it makes impossible downstream.
Read the full brief →
Deep dive — 03
Six problems, six machines: why MARS is not one insight.
The architectural consequence of designing around six failure modes instead of one insight.
Read the full brief →
Every module ships in the complete MARS package.
One price. Eleven workbooks, three TradingView indicators, and the full manual library — $497.