Neutral Market Corner

"Helled Speculation" Two ice cream sellers come to your house. One sells chocolate at $6 and the other candy at $3. Your trader yells like crazy: How can you charge so much for a chocolate cone if your competitor gives it a lot cheaper?! "Guideline" Stage A "Let me teach you how to run a business with fair prices - I bet your price will go down, so I sell in short 10 contracts of your expensive chocolate ice cream!" Stage B "Let me show you how to succeed - I bet your price will go up, so I'll take away 25 contracts of your candy ice cream, buddy!" "Spreading" "I want fair prices - sell in short 10 contracts of expensive chocolate while I buy 10 from the cheap competitor." "Neuter Chocolate Strategy" "Let me sell 10 contracts of chocolate and buy 25 of caramel. Balanced, without direction." "Directal that Sale Mal" Stage A "Wait, wait! People keep paying more for chocolate?" Seriously chocolate?" Stage B "Let me stay!" Make something to attract more customers!"

Summary of current methods to achieve "Neutrality" It will be published after Basic Statistical Technical No maximum, no minimum. It must be Bose!

This is gonna get interesting, Dave.

Summary of current methods to achieve "neutrality" Fundamental classes of assets with high geopolitical correlation: European Union, oil exporting countries, Oceania, South America, etc. High intramarket correlation: calendar spreads, spot/future differences, options on spot, HFT with temporary delay, etc. High intermarket correlation: assets listed in USD, companies of the same sector, same target audience, etc. Triangular correlation: EURUSD-EURGBP-GBPUSD, USDCAD-CADJPY-USDJPY, etc. Portfolio of long/short shares Arbitration by mergers Technical N/D Statistical Distance measures: standardized deviation, co-integration, linear regression, beta coefficient, etc. Direction/force measures: correlation coefficients (Pearson, Spearman, Kendall

As I am not a fan of standard deviation or assuming normal distribution (more parameters = more overadjustment)

Often they detect false relationships because of a "third element", usually the USD. For example, USD/CAD and WTI oil seemed correlated, but in reality they only followed the same strength as the dollar. They were co-integrated before, but no longer, although many still believe it. I also don't believe that trio like EURUSD-EURGBP-GBPUSD are fundamentally correlated. They are only mechanically bound by the EURGBP formula = EURUSD/ GBPUSD. No greed. No fear. Only mathematics.

That said, I'm very interested in how you handle positions without using standard deviation or quantiles. I hear you. No greed. No fear. Just math.

Yes, I agree. Beware of spurious correlations when there are common dividers; but on the other hand, if you find a correlated or co-integrated pair with no common denominator, would you operate without more filters? (It usually happens more with co-integration.

Great thread... thanks for opening it. As for the coverage ratio, I used a very simple formula that worked well: Ratio Par A = (ADR Par B * value pip B

Thanks for contributing. I know that idea but I did not follow it because the calculation window is something subjective. I think that formula is a kind of Beta adapted to trading pairs. I saw it in "Trading Spreads and Seasonals", if my memory does not fail. I am glad to know that it has worked using only one parameter. No maximums, no minimums. It must be Bose!