How can I get the linear regression of analyst EPS estimates? Do I need to use FHist to get the data points for LR to work on?
Few things first:
FHist(“NextFYEPSMean”,1) in the fundamental chart is not adjusting for all future splits which is why you see the jump when FTNT split 5:1. It’s a problem specific to the fundamental chart and FHist. We should have this fixed soon
However You should not use NextFYEPSMean for regressions. It’s a point in time factor and you could be regressing estimates from different years, for example soon after Q4 is announced.
What you should use is ConsEstMean(#SALEY, 1, weeksAgo) and use CTR for weeksAgo. This will always point to the same fiscal year. The “1” is for Next, “0” is for current.
But there’s another problem. The weeksAgo parameter is currently limited to 13, so you can only regress the last 13 values of the annual estimate. Not sure why we limited it to 13. We probably did not envision your use case.
How many data points do you want to regress? How far back? If you only need around 3 months then it works.
Excellent! I had forgotten about ConsEstMean. I think the last 13 weeks will be OK. I’m just starting to look at this.
And I think I understand the charting issue.
Thanks!
Probably need more than 13 weeks. That’s just one QTR and it looks like significant changes are done on a quarterly basis. Something that covers a year would be more helpful.
Yeah, 13 is useless . It’s got this strange harmonic shape with values close to 0 at the beginning and end of a quarter. It’s the correct shape though: with every quarter announcement the sales estimate gets revised and the full brunt of the revision affects the slope in the middle of the quarter.
For what you want , the slope of the annual estimate , you need 1 year minimum. I’ll add a task to get this changed.
Thank you! Much appreciated!
The weekAgo parameter limit for the Consensus Estimate Functions is now max of 100 from the previous 13.
Note that using a value of 30 or more for interims can result in many NA’s depending on what month you are on.