Built-in (hipster) FSP set

[goodboy]

piker is supposed to about minimalism, speed, and power, ideally that means our built in "indicator" (what we refer to as fsps) set should be small, objective and as hip (in the alpha generating or at least loss preventioning senses) as possible.

Most other tina platforms that retail traders get access to are filled with mostly useless or at least statistically useless "indicators" and you're stuck digging into whatever terrible scripting language the platform offers to get anything of value.

The base that seems sensible without any research:

• dollar volume (since it's the only measure of volume you can compare across instruments)

• vwap (we've already got a first draft of this without doing proper tick aggregation i.e we're using OHL3 for typical price currently)

• rsi (though this one is debatable is does track non-mean momentum which is somewhat believable as an uncommon case)

• supertrend

  • that gist account also has a slew of amibroker scripts
  • blog post from 2020 all about it
  • another blog post with LinReg
  • a tutorial on settings adjustments
  • the tinaview scripts category

• trend calculus via code we have yet to push for https://github.com/bytesumo/TrendCalculus/issues/1

  • see further discussion / resources in https://github.com/pikers/piker/issues/103

• possibly a set of other non-arithmetic means

  • harmonic
  • geometric
  • notes on inequality with the arithmetic mean used by tinas
  • other non-std weightings for moving averages

---

Some projects to dig into and research (this list should be ongoingn):

• LPPLS modes for "bubble detection"
  • checkout lppls for py pkg
• other potential hipster smoothing techniques using tsmoothie