<@U06MW6C1QM7>: did you see see article? I think i...
# communicationsm
Max
10/02/2024, 6:49 PM@Tracy Teal: did you see see article? I think its great.
👀 1
Max
10/03/2024, 5:57 PMI think we should put a disclaimer like:
It is important to mention that for many financial time series, the best estimate for the price is often a random walk model, meaning that the best forecast for tomorrow’s price is today’s price. Nixtla’s StatsForecast library allows you to easily implement this model and variations.
Max
10/03/2024, 5:57 PMWe shouldn’t create the idea that one can easily predict stocks.
Max
10/03/2024, 5:58 PM@Nikhil Gupta had some further comments that might be interesting to consider.
Max
10/03/2024, 5:59 PMI think the comments show the concerns
Max
10/03/2024, 5:59 PMI’d be more interested in seeing the results for returns rather than prices, as stock prices are generally non-stationary. Additionally, it would be valuable to benchmark these predictions against the historical average of these returns, for context. While I have a lot of respect for the Nixtla team and their work, this particular post feels more like clickbait to me. It borders on charlatanism, which is disappointing given their usual standard of rigor.
n
Nikhil Gupta
10/03/2024, 6:00 PMThat was the same concern I expressed to Max yesterday (1:1).
Nikhil Gupta
10/03/2024, 6:01 PMI did not think we would be posting this on LI though, or else I would have definitely spoken up.
Nikhil Gupta
10/03/2024, 6:04 PMHere is the background of our conversation from yesterday for what it is worth.
Nikhil Gupta
10/03/2024, 6:13 PMShould we take down the post?
a
azul (she/her) (nixtla)
10/03/2024, 6:13 PMthank you @Tracy Teal @Max @Nikhil Gupta! we've deleted the post, we think it might be good to iterate the text a little bit to reflect that stocks are not forecastable in general.
t
Tracy Teal
10/03/2024, 6:14 PMThanks all! Yes, sorry, I didn't see the discussion with Nikhil and Max. That's important context.
n
Nikhil Gupta
10/03/2024, 6:15 PMNo worries Tracy. It was actually 1:1. I should have maybe discussed in the same group where Max originally posted.
t
Tracy Teal
10/03/2024, 6:16 PMYes, thanks, that is helpful. From Max's post, I thought we wanted it shared.
m
Max
10/03/2024, 7:58 PMYes, sorry @Tracy Teal! I meant the post was great in the sense that it showed how to use timegpt and even included screenshots of the dashboard. It should have been clearer about our general position regarding stocks forecasting.
Max
10/03/2024, 8:00 PMAlso, this is a great example of why public conversations are better than private ones. I also understand that maybe @Nikhil Gupta thought I was really excited and did not want to shame me hahahah
😜 1