04 Oct

A short and amusing story for the weekend.

Last week I was in Rome. Before leaving, I read the guidelines on how to identify a restaurant that is a tourist trap: looks good on the outside - signs and a menu in English, the classic extroverted markers (grandma in the kitchen), suitable for describing a familiar pattern. Finally you get real food but not authentic and you had to pay too much.

I made the following forecast and analysis to FB after the trade closed yesterday.

In FB, as in a tourist restaurant,  the pattern for reversal looks familiar, the technical lines look familiar, the grandmother is in (a strong company overall), and the signs are shouting "buy".

I ran the weekly forecast model (one out of 5 layers) . This time with two lines, a standard in purple, plus a yellow line which  highlights the changes.
- First image shows the forecast, without the price, starting at the beginning of September. The pink right side is the forecast side. Just a reminder, the forecast shows a trend, not exact price levels.
- Second picture, same forecast and the price is updated on the graph. We can see a nice price adjustment. And we cn see that it may rise next week, for a local high and then drop again to mid Oct.
- Third photo, I left the September lines, moved the algorithm's calculation point to the end, which is October 3, and ran again. This time the lines remained in the same colors but thicker so we could distinguish between the new lines created 3 Oct and the older lines created on Sept. 

What we got - the algorithm keeps the direction even when it contains the price of the past month - consistency is very important in neural-network calculations.
However, beyond that, the algorithm clearly shows - hey friends, there may be a tourist trap here! A small increase next week and then a strong decline the following week. Be careful not to overpay this time.

Have a good weekend and hope you enjoyed it.

This should not be considered
as recommendation  to buy or sell any price instrument.

* The email will not be published on the website.