looks like we had a cvr3 buy signal at close today. has anyone backtested this? http://biz.yahoo.com/tm/030728/10468_3.html CVR 3 Signal Now, let's move on to the CVR 3 signal and then we'll talk about entry and exit. My CVR 3 signal was co-created with Dave Landry. What we found is that when the VIX moved 10% away from its 10-day moving average, it identified a market that had been "stretched too far" and was likely to reverse. In fact, over the last nine years, it has correctly predicted a two- to three-day reversal better than 68% of the time. The rules for the CVR 3 are as follows: For Buys: Today, the low of the VIX must be above its 10-day moving average. Today, the VIX must close at least 10% above its 10-day moving average. If rules 1 and 2 are met, buy the market on the close. Exit (on the close) the day the VIX trades (intraday) below yesterday's 10-day moving average (reversion to the man). Or exit within two to four days.
The signal you described involves the looking at the low of the VIX for the day. Since not all stocks and options open at the same time, does this mean that the first 15 minutes or 30 minutes of VIX data is spurious? If this is true, does this mean if the low of the VIX is set in first 30 minutes, then the backtested results are also spurious? What do you think? -- ITZ
Doesn't say the magnitude of the "reversal". Wanna bet that there's probably some kind of "reversal" 2 to 3 days from almost any randomly picked day 68% of the time? I've been watching the CVRs for a long time. Any one kicking off isn't to meaningful. If you get several at the same time, then you're more likely to have a meaningful result. In this case, it's more than CVR3. CVR1 and CVR2 hit yesterday too. So we could be due for at least a short term bounce by Friday.
It could be the specialist getting long and short. They all trade differently so maybe you're on to VIX's style.
ITZ, you bring up an interesting point. I would imagine that enough stocks are open to give a representative sample so that the calculation of the (opening) VIX would be fairly accurate (based on the sample). I don't pay much attention to the intra-day action, but again, your point is valid. For me, best way to use the VIX is to look for market reversals when the VIX is stretched away from its moving average. In other words, a CVR III. The advantage of the CVRIII is that it tends to lead the market. The disadvantage of the CVR III is that it tends to lead the market. Therefore, I use stretched VIX conditions as more of a "get ready" type of indicator. Then, look for some other catalyst, usually price confirmation. The CVRIII-modified is somewhat more accurate but occurs less frequently due to intra-day reversals in price. I also like to look at relative highs/lows. In fact, if you bought on a 10-day VIX high and exited on a 5-day low, it actually "tests out". If interested, email me and I can parse out the slides from my video on this with the rules and results. Keep in mind that the VIX, like anything isn't perfect. Dave
FYI - the VIX doesn't start being reported by CBOE until 15 minutes after the market opens - 9:45am ET By then the S&P 100 stocks and the OEX options (the basis for the VIX) are all trading
Thanks Arch for your knowledgeable reply. By the way, using the RSI also helps give a little bit more info when used with the CVR.