# Work out the right equation to calculate an output (financial data)

## Contest Brief

I need to get the right equation to calculate what is known as the "VXX roll yield".
Within the attached spreadsheet, I have all the required data to calculate this. I also have the target output (e.g. I have pre-calculated values for the VXX weekly roll yield (WRY)).
The calculation for the Daily Roll Yield (DRY) can be found here:

You will need to login to seekingalpha (it is a free/easy signup process) to see the whole story. This will give you some background about VXX and what the roll yield means. However, the equation to calculate the DRY from the above article is as follows:

DRY= ( (V2-Vs) - (n/n0)*(V2-V1) ) / ( n*(V2-V1)-n0*V2) )

Where:
n= number trading days remaining in the monthly cycle
n0= total number of trading days between the current monthly VIX futures settlement dates
V1 = Front month price
V2 = Second front month price
Vs = Spot VIX price

In the attached spreadsheet, I have tried to calculate the WRY using this formula. However, it gives a different output to the pre-calculated target values. If we plot the two against each other, we can see that they are correlated (84%). Therefore, my calculation is not quite correct.

I will award the prize to whoever can get the calculation correct so that it sufficiently predicts the expected output with 99% (e.g. r=0.99) or greater correlation.

One problem with my equation is that the above article on SeekingAlpha states that the equation is only an “approximation”. Therefore, to solve the problem you may need to use google to find the correct way to calculate the weekly roll yield of VXX. You will also need to problem solve and use your brain! Good luck!

Winner

## Employer Feedback

“Clever guy. He had a novel approach to a difficult challenge, and also communicated well.”

AntiMatt3r, United Kingdom.

## Public Clarification Board

• ###### marcostango7
• 1 year ago

R2 = 0.9955
The equation for the expected value is quite long, but the correlation coefficient is incredibly high. It is precise for those 1659 values, but of course I cannot tell if it will work for any set of values.

• 1 year ago
1. Contest Holder
• 1 year ago

OK I have extended the contest. Post an entry, and then it would be good to privately show me your equation, and we can both run it on more data.

• 1 year ago
2. ###### marcostango7
• 1 year ago

I posted an entry, if you want send me a PM and I can send you the equation.

• 1 year ago
• Contest Holder
• 1 year ago

• 1 year ago
1. ###### marcostango7
• 1 year ago

Coefficient of determination (R^2) = 0.9955
Correlation Coefficient (r) = 0.9978

• 1 year ago
• Contest Holder
• 1 year ago

What is your R2 or correlation with the expected output? Yes, I could extend, but it needs to be worthwhile...

• 1 year ago
• ###### marcostango7
• 1 year ago

I found the formula now, but it's too late to post it. If no one else has a better option, I can give you my one.

• 1 year ago
• ###### marcostango7
• 1 year ago

Is there any chance you could give us more time? I found a formula that works for the first 100 values, give me a bit of time for the rest if you can. Thank you.

• 1 year ago
• Contest Holder
• 1 year ago

The pre-calculated I did not work out. These were already done for me. I'm trying to work out how to get these values.

• 1 year ago
• ###### arafatarr
• 1 year ago

Can you please tell me how did you pre-calculated values for the VXX weekly roll yield (WRY) in column H?

• 1 year ago

## How to get started with contests

• Post Your Contest Quick and easy

• Get Tons of Entries From around the world