This week Sentient did one of those things I dislike. In WFC, in an unpinned analysis, it changed the location of a 54m trough from OCT2017 to OCT2013. That’s a big deal, and causes differences in CML estimates.
So, I decided to review all my cases and share them for a possible discussion.
I have my data from 1990 with:
1- DEFAULT unpinned model: Hurst nominal lengths
2- MY WFC model with 54m trough @ OCT2013
3- MY WFC model with 54m trough @ OCT2014
4- MY WFC model with 54m trough @ OCT2017
These are the graphs:
1- The DEFAULT model is placing an 18yr trough in MAY2017, which seems to me to be completely off. Even so, Sentient thinks we are close to an 18m trough and the CML expects a ‘double bottom’, with troughs in JAN and MAR2019.
2- MY WFC model with the 54m trough @ OCT2013 is expecting an 18m trough, that is simultaneously a 54m and a 9yr trough. So, the CML is bullish from about now to JUN2019 (where an expected 18m peak lies). Note that Sentient places the 54m trough here (OCT2013) in the unpinned analysis.
3- MY WFC model with the 54m trough @ OCT2014 is expecting a very similar situation as for the previous model, with a possible 9yr trough in the making. CML with same bullish estimates. This means that if the 54m trough is either in OCT2013 or OCT2014 makes little difference to Sentient.
4- MY WFC model with the 54m trough @ OCT2017 is expecting currently a 20wk trough. CML estimates a rise till FEB2019.
I think I am covering about all ‘predictable’ cases. Comments welcome.
During the first semester of 2019 we shall see if Sentient is indeed right…