Surprise% anomaly

Surprise%Q1 has very significant deviations from formula “(EPSActual(0,QTR)-EPSEst(0,QTR))/Abs(EPSEst(0,QTR))*100”

Any idea what is causing this?

Total cases in which Surprise%Q1 != (EPSActual(0,QTR)-EPSEst(0,QTR))/Abs(EPSEst(0,QTR))*100:

Today (2020-07-17): 32
6 months ago: 94
1 year ago: 148
2 years ago: 230
5 years ago: 479
10 years ago: 759
15 years ago: 687
18 years ago: 390
19 years ago: 1
20 years ago: 0
21 years ago: 0


This is primarily due to subsequent stock splits. See https://trello.com/c/NB7YKYXt/46-historical-estimates-on-factset-behave-differently-from-estimates-on-compustat-for-stocks-with-subsequent-price-adjustements-due . . . We hope to have a fix for this soon.

Unfortunately, problem persists as both formulas give identical results. They should give identical results, shouldn’t they?

Thank you for your quick response.


Ok, thank you Yuval!

Best to check trello or will you report back in this thread when this is resolved?

Any idea how this difference can be explained if Surprise%Q1 is calculated by P123 using the same formula as we do? Or is it a different formula?

Thanks again.

I’m assuming that one or more of these items is adjusted and one or more is unadjusted, but we need to look into this to see exactly what’s happening. I’ll try to remember to respond in this thread once the issue has been resolved.

Thanks Yuval!