Can somebody explain (and give examples) of "signals" in quant investing? What are those? What does this word mean?
6
-
2$\begingroup$ This borders on opinion, but in short 'signal' is the information that allows you to make successful investment decisions. Happy to take note if others wish to offer successful personal examples. :) $\endgroup$– ChrisCommented Sep 23, 2020 at 4:07
-
3$\begingroup$ In Quant investing buying/selling is (usually) automated. In a narrow sense the Buy Signal is a boolean variable which is 1 if the algorithm wants to buy a certain asset and 0 if it does not. And similarly for the Sell Signal. But people do use the term in a slightly broader sense to also refer to the formula/code that calculates that 0/1 from observable information or to that information itself. $\endgroup$– nbbo2Commented Sep 23, 2020 at 9:36
-
1$\begingroup$ @noob2: What you said was great but is it also the case that besides 0 and 1, the number could be in the contnuum so that 0.8 means strong buy and 0.1 means not so strong buy. So, my point is that the number can also be between 0 and 1, AFAIK. I don't use this approach so I'm not positive but it makes sense because then you can start pooling signals from different strategies together and start ranking them etc and maybe do some optimization with them or whatever. $\endgroup$– mark leedsCommented Sep 24, 2020 at 3:03
-
$\begingroup$ Agree that you could have a continuous signal, it is not always a discrete 0/1. Thanks. $\endgroup$– nbbo2Commented Sep 24, 2020 at 9:28
-
$\begingroup$ Thanks! But what would be for example a "momentum signal"? $\endgroup$– QwertyCommented Sep 24, 2020 at 13:53
|
Show 1 more comment
1 Answer
$\begingroup$
$\endgroup$
As people in the comments noted, signal broadly refers to a trigger variable that denotes an investment decision. This is normally a boolean variable (i.e. 0 or 1) but could be continuous (0 to 1) or any other range (e.g. -1/0/1 sell/hold/buy), depending on what your execution algo might dictate.
Just wanted to add that this terminology comes from the general field of signal processing, where, broadly speaking, you take noisy information and translate it into meaningful information. In this case, you are taking market data and transforming it into a buy/sell decision.
More specifically on signal processing though, Digital Signal Processing is a somewhat popular subfield for quantitative finance, and there are some interesting work done on this.