Question about implementation of "secure f"

[posman_]

Hi

I'm in the process of implementing the calculation of the "secure f" but there is something I don't understand.

I've already read the article explaining the algorithm and I understand it almost entirely. But there is still one part that is not entirely clear.

The article I'm talking about is "Secure Fractional Money Management" by L. Zamansky, Ph.D., and D. Stendahl published in Stock & Commodities magazine https://store.traders.com/-v16-c07-051secu-pdf.html (paid version)

The algorithm is simple, and its implementation should be simple too. Here is the algorithm described in the article:

Step 1 Calculate optimal f
Step 2 Set f = optimal f
  If the acceptable maximum drawdown is greater than or equal to the historical maximum drawdown, then stop; otherwise, go to step 3
Step 3 f = f - delta
  If f > 0, then
    * Calculate MDD(f) - the maximum drawdown for the f portion of the capital to be invested in every trade using the information about trades and prices
      If MDD(f) < D, then calculate TWR(f, D) and store the values ​​of f and TWR(f, D)    
    * Return to the beginning of step 3

  If f < or = 0, then
    * Find the value of f such that corresponds to the maximum of all TWR(f, D) stored. This value is the secure f
    * Stop

The part I don't understand is in step 3, Calculate MDD(f). I must calculate the drawdown but using a different "f". What I don't understand is, if I use the same list of historical operations, why should the drawdown be different? The drawdown is not calculated using the "f", but only the profit/loss of the operations.

I'm probably missing something that's very simple.

Has anyone had experience with successfully implementing "secure f"?

Thanks in advance

[Stanislav Korotky]

IMHO, Drawdown is calculated from Capital+Profit/Loss, so it will change if you take "f portion of the capital".