poppa

poppa

01-11-2022

01:11

INTERPRETING THE DENSITY PLOTS: A THREAD

WHAT IS A DENSITY PLOT? by definition, density plots are a representation of a distribution of numeric variables. in practice, Griswald (our bot) uses them to visualize the distribution of outcomes for an asset; usually $SPX

in short, we use density plots to visualize the distribution of probabilities for the one day and five day forward return of $SPX — and in plain English, we ask "what are the most likely outcomes for $SPX on a one (five) day timeframe?" LEARNING IS FREE:

so, how does Griswald tell us which outcomes are likely and which aren't? he sifts through 15 years worth of data then filters around ~20 days that are the most 'similar' to the previous day's close using a set of indicators. he then paints us a picture. every morning.

confusing at first, I know. the green squiggly line is the distribution of outcomes for days that are 'similar' per the indicator — observing the graph below, 'similar' days typically have a strong propensity to close +0.5% (i.e. where it is the tallest)

also, that little hump on the left at -1.5% is important as well, but lets not get ahead of ourselves.. the blue squiggly line is the distribution of outcomes for $SPX with no filter — it has a 'normal' distribution. yellow vertical line = median green vertical line = mean

now, the intuition.. since the blue squiggly line is what we can expect in general for SPX, and the green line is our 'forecast' — wherever the green line is above the blue line we consider our 'edge' (i.e. it is more probable per the forecast than what we usually expect of SPX)

so...we ask Griswald to make this a bit easier for us to interpret.. Griswald, can you calculate the difference between the green line and blue line? Voila — the stripe. the taller the stripe, the more edge in the outcome. additionally, it's colour corresponds to the % return

quick recap - Griswald is a bot that generates density plots for us - He filters 15 years worth of data for approximately 20 'similar' trading days as a forecast - The green squiggly line are the % return (x axis) of the 20 trading days - Taller squiggly = more probable

quick recap cont - Griswald calculates a *very* dirty version of edge for us to interpret quickly; the difference between 'what usually happens' (blue), and 'what we think will happen' (green) - The 'edge' stripe's colour is determined by the % return outcome (x axis)

if that's not confusing enough, Griswald then takes 10 different density plots based on 10 different indicators to paint us one single plot. well, two different plots. one where volatility based indicators are weighted more, and one where non-volatility based indicators are

all the same rules apply, but there are a couple additions: 1) the dots represent the 'max edge' forecasts for all the indicators (i.e. where the blue stripe is the highest for each individual plot 2) this is a collection of all the forecasts with some other summary statistics

[Forecast] — max edge forecast when combining all indicators [Numbers] — max edge forecast for all indicators [Edge] — bullshit dirty edge calculation [Probability] — bullshit probability calculation [RNG Forecast] — bottom range, top range, s.d. of range (20d lookback)

[SuMo] — similar to the RNG forecast: bottom range, middle of range, top of range

In the mornings I will sometimes tweet a GIF that is essentially a powerpoint presentation for you. each slide of the deck is the individual density plot for all the indicators. Don't worry tho fam. I will hold your hand.



Follow us on Twitter

to be informed of the latest developments and updates!


You can easily use to @tivitikothread bot for create more readable thread!
Donate 💲

You can keep this app free of charge by supporting 😊

for server charges...