[scalper4]

Does anyone have code for daily volatility bands as used by Kevin Haggerty?

[steve46]

Hello

Since you don't mention a specific platform (tradestation for instance) I assume what you really want is a mathematical formula for intraday vol?

I am not familiar with Haggerty's method but can offer my own as follows;

1. Go to IVolatility.com and get the most recent IV (Implied Vol) for the instrument you trade. If for example it is the ES contract, then it is approximately 37%....convert to .37

2. To compute a one (1) standard deviation trading range (annualized)

1 x .37 x previous day's closing price (900 for the ES contract) = 333 pts above and below that close

To obtain the intraday figure simply multiply by the square root of the number of days per year (365) shown below

square root of 1 day/365 = .0523421

333 pts x .0523421 = 17.43 pts

Indicates that the intraday 1 sd range for the ES contract should be 900 plus or minus 17.43 pts.

or

882.50 - 917.50

Rinse & repeat to obtain 1.28, 1.5 and 2.0 standard deviations and you have "approximate" intraday ranges for the ES today (Jan 02, 2009)

Here they are out to 1.5


+1.5 sd = 926.25
+1.28 sd = 922.50
+1 sd = 917.50

900 (previous day's close)

-1.5 sd = 882.42
-1.28 sd = 877.60
-1.5 sd = 873.75

Remember that its an approximation and that it "suggests" that the price series is normally distributed (it isn't)..so there are quite a few limitations to it. I wouldn't use it but there it is....

Hope it helps

Steve

[MidKnight]

Steve,

Out of curiosity, which IV are you using, the calls of puts? Or perhaps, the put IV for downside range and the call IV for the upside range? IVolatility doesn't deal to international markets so I have to try and make do with the best I can. This is an end of day spreadsheet of all the IVs for all the options from the exchange. I assume you are using ATM IVs.

My thanks in advance,
MK

[steve46]

Hello

I just read your question. I just used the IV offered by IVolatility on their free home page. Frankly a more rigorous way to do it is to take both ATM put and ATM call IV, sum them and take the average. Once again for the record, there are a lot of holes in this idea....for one, price is clearly not distributed normally, and the numbers you generate with this method are really just an approximation. At the very least I would want to take trades that occur when at least one other supporting signal is nearby (the principle of "confluence" from my thread "Ideas for Struggling Traders").

Finally, I have no idea now Mr. Haggerty is using this concept. He may have some novel way of using it that I am unaware of.

Hope this helps
Steve

Edit

Thinking about this, if one wanted to get a more accurate set of numbers (and they had some math background) the way to do it would probably be to obtain a weighted average IV or estimator as is used in GARCH 1,1.

[MidKnight]

Thanks Steve,

I will explore this further. I'm wondering if tracking the spread between the advertised price vs the theoretical price of the ATM calls/puts can offer any directional insights. I will explore it further.

All my best,
MK

[Fedeo]

Hi,
I am quite familiar with the Haggerty methods, I can say that the formula is exactly the same as the one highlighted by Steve. I can just add that Haggerty normal use that formula in the first 30-60 minutes of trading to "evaluate" an excess on the up/down side due to some news. After getting those levels he expects then for some patterns to happen to trigger the entry. Honestly I would say nothing really new if compared to most of the methods and strategy that can be found on TL.

Fedeo

[scalper4]

Thanks for the reply Steve,
I was looking to see if anybody had coded up an automatic indicator that pulled the data down and did the calculations. I have seen one for sale on ebay but it is expensive and I have seen other people use them. I use tradestation and wondered if anyone had the code.

[jeanieg]

I have the one from EBAY and it works great. Sometimes the support and resistance areas are "spooky" in their accuracy. Don't know if the code is out there but the package I purchased works for me.