Testing on XAUUSD (Gold)? about code HOT 3 CLOSED

Thank you for your interest in our work. Could you please explain a bit what you mean by "failed tests" for XAUUSD?

from code.

Apologies, I mean it blew the account in 2011 as a result of too high a drawdown. I started with $1m. The largest drawdown using FX was around 1.0 - 1.8% for a similar back test period. Gold started at around $600 in 2007 then was around $1,900 in 2011 - this is a possible cause as the change in price was simply too much. It should be noted that runners in my programming end deltas at 100%. Is it intended that delta up/down exceed 100%?

from code.

The process of skewing (Section 3.6 in the "Alpha Engine" paper, i.e., dStarUp/Down in the code) is the mechanism that can increase or decrease thresholds. It is a function of the inventory. A simple implementation is to define a minimum (maximum) cumulative position size and then increase (decrease) the deltas of newly initialized coastline traders (e.g., double or halve). This should be done separately for long and short positions.

Regarding the XAU/USD, we have actually seen very good results for this instrument for a time period we tested between early 2000 and 2013. Indeed, even during problematic times (S&p500 quant meltdown, financial crisis, and flash crash; de-pegging of Swiss Franc; quantitative easing USD/JPY) the models performed well...

from code.