Detecting stock prices that are easy to go up and down in a short period

Question

I'm trying to use python to detect if the price of a stock is easy to go up and down in a short period.

The yellow line in the following pictures are the stock prices:

pic1

pic2

The above 2 stock prices are considered to be easy to go up and down for me.

The following stock prices are not.

pic3

pic4

pic5

I have all deal prices of the stocks.

I've tried normalize the stock price then calculate its standard deviation.

import pandas as pd

today = date.today().strftime('%Y-%m-%d')
ticks = api.ticks(   # api is an object from shioaji API
    contract=api.Contracts.Stocks[ticker_code], 
    date=today
)
# The type of ticks.close is "list"
pclose = pd.Series(ticks.close)
# normalized_std is used to decide if the stock price is easy to go up and down in a short period.
normalized_std = (pclose/pclose.mean()).std(ddof=0)

But the result is not good.

For example, pic1's normalized_std is 0.013441217877894908, pic2's normalized_std is 0.011078299230201987. But pic3's normalized_std is 0.02169184908826346, and pic5's normalized_std is 0.0014295901932196125.

So it's hard to decide if the price is too easy to go and down by using normalized_std variable.

Anyone knows if there's a method to detect this kind of stock price patterns?

Can machine learning be applied to solve this problem?

Answer 1

Just an idea, but it looks like you want a lot of absolute movement, with little relative movement. For example down $1.10, up 1, down $0.90 up $1.10

Absolute movement would be the sum of the absolute value of the moves: 1.10 +1+0.9+1.10 =$4.10
Relative movement would be the sum of the moves: -1.10 + 1 - 0.90 + 1.10 = 0.10

A fancier method, if you have access to real-time quotes, is to calculate the price of a short expiration straddle (the price of the at the money put and call), it is good indicator of expected movement.