I originally gave Julius this prompt:
Hello, I am an intraday forex trader, and I rely on daily currency strength reviews to determine trading opportunities. Currently, I receive strength scores for the 8 major world currencies from an associate, but the delivery is inconsistent, and he has not shared his calculation method.
I have gathered 17 days of his strength scores, all provided in the mornings, so his calculations are likely based on data from either the same morning or the prior day. His method involves all 28 currency pairs, and I believe he reviews each pair individually, possibly considering indicators like RSI or similar metrics.
Objective: I aim to reverse-engineer his strength formula to achieve a similarly ranked strength score for each currency daily, with an additional goal of a root mean square error (RMSE) of less than 5 between my calculated scores and his.
Resources: I can provide an API connection via TraderMade.com, granting access to historical data for all 28 currency pairs over the 17-day period.
Key Priorities:
1. Replication of Ranking - First, identify a scoring formula that mirrors the strength ranking order of my associate’s daily scores for each currency.
2. Relative Distribution - Ensure the relative strength distribution between currencies closely matches his scores each day.
3. Accuracy Target - Finally, fine-tune the formula to achieve an RMSE score below 5.
Is this clear, or are there any additional details you need to proceed?
We’ve been going back and forth trying to work this out. However, we keep ending up with it ‘running code’ and just getting stuck there. Is this normal? Is it just doing something really complex that takes a lot of time? Any help would be hugely appreciated.