I'm relatively a beginner and have been researching trading methodologies for a few months.
I had a weak hunch that bounces from highly oversold situations might tend to happen majorly through overnight up-gaps rather than extended gradual rise during trading hours. Right or not, I tried to quantify the hunch through the following test -
As the test sample, I took Nifty Futures EOD data from May 2008 to Nov 2011. I wanted to observe the behaviour of opening gaps under the following situations -
1. Overall
2. During Oversold Days (Taken as days where the Closing Price of the day was below by more than 5% from the 50 DMA)
3. During Overbought Days (Taken as days where the Closing Price of the day was above by more than 5% from the 50 DMA)
4. High VIX days (Taken as days where the closing value of VIX was 15% above the 50 DMA of closing VIX values)
The following were the observations -
No. of Days in the sample - 865
Average Overall Absolute Opening Gap Size = 0.7%
Avg Absolute Opening Gap Size on Oversold Days = 1.3%
Avg Absolute Opening Gap Size on Overbought Days = 0.7%
Avg Absolute Opening Gap Size on High VIX Days = 0.9%
% Days when Absolute Opening Gap > 1% = 22%
% Oversold Days when Absolute Opening Gap > 1% = 44%
% Overbought Days when Absolute Op Gap > 1% = 19%
% High VIX Days when Absolute Op Gap > 1% = 38%
Basically, the absolute opening gap tends to be a lot higher than average, under oversold conditions, than in normal conditions. Almost 1 out of every 2 days in oversold conditions has an absolute opening gap of more than 1%.
There is some, but not an exploitable directional bias (i.e. whether the overnight gap was positive or negative) to these overnight gaps under oversold conditions. Also, there is some, but not a complete overlap in oversold days and high VIX days.
The idea that I wanted to test based on these numbers was that of buying a Nifty Options Strangle at close of oversold days, and selling it on the next day open. The assumption being, that the loss in value of the strangle on days when there isn't a big gap will be more than offset by the profit in strangle values on days with a big gap. (Assuming that the value of the strangle should increase after a big gap.) But, with high VIX also present in many oversold cases, the option prices are already quite inflated for the long strangle to yield much benefit. Can this be tested?
The excel sheet with the data used to calculate these numbers is uploaded here - http://www.4shared.com/file/aeYsjR5J/QE12_-_Gaps_under_conditions.html
I'm looking for suggestions on how a trading idea based on these numbers can be backtested. Or, in what direction can this research be taken further.
Thanks
I had a weak hunch that bounces from highly oversold situations might tend to happen majorly through overnight up-gaps rather than extended gradual rise during trading hours. Right or not, I tried to quantify the hunch through the following test -
As the test sample, I took Nifty Futures EOD data from May 2008 to Nov 2011. I wanted to observe the behaviour of opening gaps under the following situations -
1. Overall
2. During Oversold Days (Taken as days where the Closing Price of the day was below by more than 5% from the 50 DMA)
3. During Overbought Days (Taken as days where the Closing Price of the day was above by more than 5% from the 50 DMA)
4. High VIX days (Taken as days where the closing value of VIX was 15% above the 50 DMA of closing VIX values)
The following were the observations -
No. of Days in the sample - 865
Average Overall Absolute Opening Gap Size = 0.7%
Avg Absolute Opening Gap Size on Oversold Days = 1.3%
Avg Absolute Opening Gap Size on Overbought Days = 0.7%
Avg Absolute Opening Gap Size on High VIX Days = 0.9%
% Days when Absolute Opening Gap > 1% = 22%
% Oversold Days when Absolute Opening Gap > 1% = 44%
% Overbought Days when Absolute Op Gap > 1% = 19%
% High VIX Days when Absolute Op Gap > 1% = 38%
Basically, the absolute opening gap tends to be a lot higher than average, under oversold conditions, than in normal conditions. Almost 1 out of every 2 days in oversold conditions has an absolute opening gap of more than 1%.
There is some, but not an exploitable directional bias (i.e. whether the overnight gap was positive or negative) to these overnight gaps under oversold conditions. Also, there is some, but not a complete overlap in oversold days and high VIX days.
The idea that I wanted to test based on these numbers was that of buying a Nifty Options Strangle at close of oversold days, and selling it on the next day open. The assumption being, that the loss in value of the strangle on days when there isn't a big gap will be more than offset by the profit in strangle values on days with a big gap. (Assuming that the value of the strangle should increase after a big gap.) But, with high VIX also present in many oversold cases, the option prices are already quite inflated for the long strangle to yield much benefit. Can this be tested?
The excel sheet with the data used to calculate these numbers is uploaded here - http://www.4shared.com/file/aeYsjR5J/QE12_-_Gaps_under_conditions.html
I'm looking for suggestions on how a trading idea based on these numbers can be backtested. Or, in what direction can this research be taken further.
Thanks
