Historical vs Implied spread

kv1289 · Sep 13, 2019

I have seen many traders use the spread between historical and implied volatility as a factor when entering a trade.

Does this spread actually have any predictive power?
If so in what context (ex. If historical higher than implied does that phenomenon continue? Does implied usually catch up to historical?)
Has there been any research done on this topic?
If I wanted to test this what would be some ways I could measure it's effectiveness?
Lastly, I've observed many charting packages plot implied and historical in line with each other, however we don't know the historical for today until some point in the future. Is it appropriate to view them in line or does there need to be an offset applied? (ex. The 5D Hist plotted today is backwards looking which the implied plotted today is forwards looking)

Thanks!

BlueWaterSailor · Sep 13, 2019

kv1289 said:
I have seen many traders use the spread between historical and implied volatility as a factor when entering a trade.

The first indicator I ever rewrote (because I was frankly offended by the idiotic premise behind it) was a "Professional options trader" gadget that used HV/IV to display "trade zones". They were using historical volatility - note that this is not the same thing as historical implied volatility - as if the two were in any way related. There are probably hundreds if not thousands of traders using that "indicator" today.

Does this spread actually have any predictive power?

Van K. Tharp, in his "Trade Your Way to Financial Freedom", cites having performed an experiment in which he would enter trades based on the flip of a coin and focused solely on his exits; the outcome was a positive P&L. I seriously doubt that the "HV/IV" voodoo has any better predictive power than that coin flip.

If I wanted to test this what would be some ways I could measure it's effectiveness?

Backtest it, then forward test it.

krugman25 · Sep 13, 2019

Implied volatility regularly overstates realized volatility which is the whole idea behind options selling strategies. There are so many angles to look at when talking about this topic, but it is too late for me to get into any of them so I will leave you with a few good videos with historical studies.

https://www.tastytrade.com/tt/shows...orical-versus-implied-volatilities-03-03-2015

https://www.tastytrade.com/tt/shows...des/the-chance-of-being-overstated-08-20-2019

Robert Morse · Sep 14, 2019

kv1289 said:
I have seen many traders use the spread between historical and implied volatility as a factor when entering a trade.

Does this spread actually have any predictive power?

If so in what context (ex. If historical higher than implied does that phenomenon continue? Does implied usually catch up to historical?)

Has there been any research done on this topic?

If I wanted to test this what would be some ways I could measure it's effectiveness?

Lastly, I've observed many charting packages plot implied and historical in line with each other, however we don't know the historical for today until some point in the future. Is it appropriate to view them in line or does there need to be an offset applied? (ex. The 5D Hist plotted today is backwards looking which the implied plotted today is forwards looking)

Thanks!

In my experience, when IV is higher than HV, there is a reason for it, E.G. Pending event or market conditions.
To answer your question, yes, IV will generally return to HV but it might mean that the stock had a large move 1st or the event past where there was that expectation or market conditions give rise to expectations of larger than normal moves. IMO, not a great way for a retail account to make money.

That said, watching those values has value when you add that to a thesis that the stock will move in or out of a range or not move at all. If you have some type of predictive model for this and add your expectation of what option prices will do over a time period that should work well over time. Then you can choose options that best meet your needs including the changes in IV that you expect.

To add to that, I have seen many instances where high vol was a buy and low vol was a sell. By themselves, without a strategy besides just that comparison, it is hard to make money consistently.

wildchild · Sep 14, 2019

kv1289 said:
I have seen many traders use the spread between historical and implied volatility as a factor when entering a trade.

Does this spread actually have any predictive power?

If so in what context (ex. If historical higher than implied does that phenomenon continue? Does implied usually catch up to historical?)

Has there been any research done on this topic?

If I wanted to test this what would be some ways I could measure it's effectiveness?

Lastly, I've observed many charting packages plot implied and historical in line with each other, however we don't know the historical for today until some point in the future. Is it appropriate to view them in line or does there need to be an offset applied? (ex. The 5D Hist plotted today is backwards looking which the implied plotted today is forwards looking)

Thanks!

It usually means something big is coming. For instance a company may be involved in a big legal case that is close to being decided. The stock itself may trade in a tight band as the decision nears causing low HV, the IV will be high because once the ruling comes out, the stock will plunge or spike depending on the ruling. Investors/traders will be hitting the options hard to either buy protection or speculate and cause an increase in option prices, thus causing IV to be high.

Thats just one example of what could cause it.

stepandfetchit · Sep 14, 2019

IMHO: Loosely speaking, IV is an option trader's best guess of Future Realized Historical Volatility. However, the word "loosely" should be a caution flag. To get a better understanding, the terms for HV and IV need to relate to the same information being sought. For example, for Implied Volatility, many people merely reference the broker supplied 30 calendar day computed volatility, which is not appropriate to relate to Historic Volatility, since it is related to a selection of OTM option premium (so is always greater than the more appropriate Implied Volatility of the ATM options)! The time frames should be correlated to the same times as well, so insure your IV (in calendar days) and your HV (in Trading days) is compensated so they reference the same trading interval AND the trading interval is the specific interval you seek (one size fits all is not appropriate). As you noted, the X-axis (if charting) need to be considered to insure the IV and HV plots are not offset in time (to make viewing clear). Personally, since I will be trading using the IIV (forward looking), I will offset the HV, to reflect the Future Realized Volatilty (same data, just shifted in time).
You can then examine specific time frames to get better feel of how well the Implied volatility considers future realized volatility. I find the IV is poor predictor of unexpected events {Duh! should be obvious}. {I only considered SPX for above in my work} BTW: HV > IV typically associated with Backwardataion.

zdreg · Sep 14, 2019

test zzl

stepandfetchit · Sep 16, 2019

Below is repost of something I posted in another forum a few months ago that may be of interest:
-------------------
I take a peak at comparing Implied Volatility to Realized Volatility. If these are "apples to apples" comparisons, one should be able to relate to success/failure of simple "cannon-launch" trading, such as IC's. No new data here, but should merely help to confirm what we already know.

For Apples to Apples, I did the following:
1) Time: Chose same time references, and relate to Calendar days instead of Trading days. This is the normal perspective of Implied Volatility.
2) Which Implied Volatility: I chose ATM Volatility instead of the Term Volatility (or VIX calculation). This results in a smaller value than the term/VIX method and is less influenced by SKEW. Specifically, observe the volatility of the strike above and the strike below the ATM, and use linear interpolation to obtain the ATM value. (note: The IV derivation derived from process documented by Steve Speer)
3) Time (fine tuned): The targeted time is computed from the nearest two available terms and interpolated if the not exact time available. -- For Realized Future Volatility, the closest match to the "# Calendar days" is used, so if target would fall on Sunday, then Monday would be used, if target falls on Saturday, then Friday would be used, else the exact # Calendar days is used.
4) Time (visual): Since we trade on the hard right edge of the graph, I use the IV (forward looking time reference), so Realized Future Volatility is really the "Realized Volatility" shifted to the "left" to allow precise verification of the Real vs the Implied values & correlate to the SPX price at that time. --

My SPX option database is currently complete back to late 2016, so I use that as starting point, and for this exercise am using 3PM Eastern time (normal closing time) data.

I plot Realized Future Volatility against Implied ATM Volatility, and provide a relative view of SPX, then replicate this plot and add a integral of IV-RFV (SPX->DiffSum on plot) to produce a curve that may approximate the shape of an Equity curve of someone trading Cannon-Launch Premium Selling trades of that duration! Note the Y-Axis labels for the plot pairs. ( the Blue plots of "SPX->DiffSum" will precede by "length" the impact on an equity curve)

Note: "length" is the input variable to the plots! (length is Calendar Days)
First pair is 14 Days: