Was just going thro QUSMA blog on Equity Returns Following Extreme VIX and William VIX FIX Movements. And couple of peoples are throwing ideas on implementing the strategy with percentile rank function. The same discussion also happening in Quantopian Community which inspired me to code the strategy for Nifty Futures in Amibroker
For Example if Percentile rank of VIX reads 98 with an 100 period lookback. It means the readings are extreme and never occurred in the last 98 bars out of 100 bars. Percentile rank will show a reading of 98-100 range if there is a sharp spike in VIX or William VIX Fix (WCF) readings. So thought of trying out a strategy on it.
The Strategy goes like this,
Buy at the next day close if either of the VIX Percentile rank(VIXrank) or William VIX FIX Percentile rank (WVFrank) reads above 98 and also either of the readings VIXrank or WVFrank reads above 99. And Exit 4 days later at the days close.
What are the data we need collect?
Indian VIX Historical data and Nifty Futures EOD data. The real challenge is VIX data starts from 2008 onwards and NSEindia website had data from Mar 2009 onwards. So Able to backtest the strategy right from 1st Jan 2010 onwards to till date.
VIX-WVF Percentile Rank – Amibroker AFL Code
4 lots of Nifty futures with 150 rs per leg as commissions + Slippages (Rs 300 Commission for two legs Buy and Sell). Rollovers charges and premium are not considered(has to be calculated manually). Another best alternative to avoid rollover is try this strategy with NiftyBEES ETF. I leave it as exercise to you people.
If we try to do only with VIXpercentile rank the results are good with same winning ratio however with a little lower net profit compared to the combined VIX-WVF percentile rank strategy. However the strategy doesnt works well with standalone WVF strategy. Combining both VIX-WVF produce optimal results.