This spreadsheet shows the implementation of the Minimum Correlation Algorithm (mincorr). The power function is a new parameter that regulates the weighting in the averaging process toward the most anti-correlated assets. Currently the default is “1″ which is essentially a rank-weighted moving average of pair-wise correlations.  A value of “0″ would imply equal weightings, values lower than 1 would imply a balance between equal and rank weighting, while values higher than 1 would more heavily skew the rank-weighting. More spreadsheets and research will be available soon.

2 thoughts on “Minimum Correlation Spreadsheet Now Available”

1. Hi! Let me congratulate you for the blog, I always read it. Today I’m asking you forr some help! I hope you can help me!
I know this is not the space for leaving questions off topic, but I didnt find other place.

I’m using excel to (trying to) test some trading strategies. I’ve already tested them on a software and I have the data.

I have all the data of the individual Profit/Loses of the trades. Also have the date when they were oppened and closed. Of course I have the ammount of contracts/stock/etc bought (or sold in short positions).

My problem is that, I’m trying excel to tell me when a trade would actually be opened. For example if a trade was opened on 10-Oct-2011 and closed in 20-Oct-2011, and I used 50% of my equity, if in theory the software shows me 3 more trades between 10 and 20 Oct. But I’d only opened 2 of them (assuming I’d use 50% of my equity in each) in the reality.

Any idea? I have this: http://screencast.com/t/7tdIWhWx

If you see, the rows must be “grouped” 2 by 2, because they are part of the same trade. 1 shows the open and 1 shows the close.

I think the solution should be something like counting the diference between the date of close of one position and the date of open of the next and if it is negative, there should go an IF function, wich returns TRUE/FALSE if the cost of trade if greater than the cost of the already opened trades + the profit or lose of previous trades.

:S this is very confusing. Thank you for your help!

2. thapelo says:

hi David,
thanks for your wonderful research. however I have one questions, there a way to add bounds (more the upper bound) to the weight? cause that will be a piece of useful additive.

thanks