Computing a rough HV estimate from limited data

earth_imperator · Nov 5, 2022

A question for mathematicians / statisticians / quants / TAs:

Say you have only the following data of a stock (just this single daily data).
It lacks the annual historical volatility (HV) data, ie. the StdDev.

Is it possible to roughly estimate a HV for this stock using the data below?

Would it be helpful having such data of max 5 consecutive days?

Code:

"quote": {
  "fiftyTwoWeekLowChange": 6.9799995,
  "fiftyTwoWeekLowChangePercent": 0.1279091,
  "fiftyTwoWeekRange": "54.57 - 164.46",
  "fiftyTwoWeekHighChange": -102.91,
  "fiftyTwoWeekHighChangePercent": -0.6257449,
  "fiftyTwoWeekLow": 54.57,
  "fiftyTwoWeekHigh": 164.46,

  "fiftyDayAverage": 75.2422,
  "fiftyDayAverageChange": -13.6922035,
  "fiftyDayAverageChangePercent": -0.18197505,

  "twoHundredDayAverage": 95.26105,
  "twoHundredDayAverageChange": -33.71105,
  "twoHundredDayAverageChangePercent": -0.35388073,

  "regularMarketChangePercent": 4.855193,
  "regularMarketPrice": 61.55,
  "regularMarketChange": 2.8499985,
  "regularMarketDayHigh": 61.88,
  "regularMarketDayRange": "58.69 - 61.88",
  "regularMarketDayLow": 58.69,
  "regularMarketPreviousClose": 58.7,
  "regularMarketOpen": 58.765,

},

globalarbtrader · Nov 5, 2022

earth_imperator said:
A question for mathematicians / statisticians / quants / TAs:

Say you have only the following data of a stock (just this single daily data).
It lacks the annual historical volatility (HV) data, ie. the StdDev.

Is it possible to roughly estimate a HV for this stock using the data below?

Would it be helpful having such data of max 5 consecutive days?

Code:

"quote": { "fiftyTwoWeekLowChange": 6.9799995, "fiftyTwoWeekLowChangePercent": 0.1279091, "fiftyTwoWeekRange": "54.57 - 164.46", "fiftyTwoWeekHighChange": -102.91, "fiftyTwoWeekHighChangePercent": -0.6257449, "fiftyTwoWeekLow": 54.57, "fiftyTwoWeekHigh": 164.46, "fiftyDayAverage": 75.2422, "fiftyDayAverageChange": -13.6922035, "fiftyDayAverageChangePercent": -0.18197505, "twoHundredDayAverage": 95.26105, "twoHundredDayAverageChange": -33.71105, "twoHundredDayAverageChangePercent": -0.35388073, "regularMarketChangePercent": 4.855193, "regularMarketPrice": 61.55, "regularMarketChange": 2.8499985, "regularMarketDayHigh": 61.88, "regularMarketDayRange": "58.69 - 61.88", "regularMarketDayLow": 58.69, "regularMarketPreviousClose": 58.7, "regularMarketOpen": 58.765, },

Yes under certain assumptions you can derive a ratio between the N day range and the N day standard deviation, closed form mathematically or with simulated data. Or measure it for similar instruments where you have this data.

Having multiple days won't help very much.

Obviously with a long enough history of daily returns you can calculate the standard deviation directly.

GAT

earth_imperator · Nov 6, 2022

globalarbtrader said:
Yes under certain assumptions you can derive a ratio between the N day range and the N day standard deviation, closed form mathematically or with simulated data. Or measure it for similar instruments where you have this data.

Thanks! Appreciate it, it's encouraging to research it further.