alt Hacker NewsI see that a lot of these are markets.
Yes, it’s hard to predict markets. Because anybody who can successfully predict markets, does so, makes money, and changes the market so their predictions lose their edge.
Time series forecasts are a lot easier if you are forecasting, say, disk use in your servers or whatnot. (By “easy” I mean you can do a simple prediction and get useful insights.)
Replies
Exactly. A lot of academics seem to not get the fundamental problem: You don't know the future.
You don't know what Trump is going to say 2 hours from now. You don't know what natural disaster is going to happen tomorrow. You don't know what war is going to break out next month.
NOTHING in your past data contains anything that can tell you these events are going to happen.
Now markets may have idiosyncratic residuals from momentum and reversion effects that you can quantitatively model and profit from, and that's a tradeable signal, but the way you do that is realizing that a certain coin is slightly biased and trade it a million times, averaging out the news shocks and recovering the residual idiosyncratic bias that you found.
Trying to forecast actual prices beyond ultra-short horizons is trying to predict those shocks, which is a fool's errand. You have a system with a signal to noise ratio of 1:100, and you're effectively trying to predict the noise instead of the signal.
Between the difficulty and value propositions for forecasting timeseries "disk usage" versus "financial markets" there are some rather relevant time series such as "company sales" or "demand for company product" that come up again, and again, and again but are neither "easy" nor "predicting-the-stock-market-hard".
It's true that markets are more _adversarial_. But there's still a lot of trouble with distribution shifts even in server metrics. As an example, our SRE team got paged a few times in the past month for traffic drops due to the World Cup. This stresses the nowcasting alert in several dimensions:
- there's no seasonal pattern to the matches, they happen sorta randomly.
- they drive increased query traffic in the hour or so before the game
- then during the game usage drops, sometimes to below "normal" depending on time of day and who's playing
So... now the accuracy of your forecasting tool depends on correctly predicting when world cup matches happen, and also who wins them!
edit: and this is just one recent example. others involve severe weather, national gameshows, earthquakes, and when you celebrate christmas.