Earlier Prop Firm Analysis Overview
What the five steps answer, what they compute on, and what they explicitly do not claim.
These pages describe the prop firm analysis as it used to run as a mode in the dashboard's Portfolio Builder. Today the Prop Firm Analysis sits in the QuantDesk sidebar and calculates differently, see Funding and Prop Firm Analysis.
The earlier prop firm analysis runs a portfolio against the rules of an evaluation account. It answers five questions, in this order:
| Step | Question |
|---|---|
| 1 · Account | Which mechanism is the calculation run against, and what are your account's numbers? |
| 2 · Compatibility | Which of your strategies can actually be traded on which model? |
| 3 · Scaling | Up to which risk factor would this history have kept the account? |
| 4 · Simulation | How often does the account hold, how long does it take, and what comes after? |
| 5 · Cost | What does a passed account cost, and does the path earn itself back? |
The data basis
Nine firms, sixty account models. Every rule is mapped by hand from its source and carries that source's wording with it. Where a source is silent, that is shown as its own state and not as a zero.
Three states are kept strictly apart, because they look identical in a table and mean the opposite:
- Documented. The source names a value, and its wording sits next to it.
- Not documented. The source says nothing. That is a statement about our reading, not about the product.
- Not applicable. This model has no such rule.
What the calculation runs on
Every number comes from one price history of your portfolio and from series reassembled out of it. That is a sample of size one.
The reassembled series are a block bootstrap: blocks of consecutive days are put back together in a new order. Whole blocks instead of single days, because a losing streak consists of consecutive days. Drawing days individually breaks up exactly the streaks that blow an account, and yields a friendlier number than reality.
Checks run at daily close. With a trailing threshold, the move between two closes is already priced in; with the daily loss limit it is not, and open positions do not exist in the series at all. A firm that marks continuously will therefore hit a breach sooner than this analysis says.
What the analysis does not claim
- No forecast. A history that would have kept a rule says nothing about whether the next one will.
- No recommendation. No model, risk factor or price variant is marked as "the right one".
- No completeness of rules. What is modelled is what the sources say. Where something is missing, the interface says so in the place where the number would otherwise be.