srinir - I am not a client of Kim's but I can confirm his strategy is valid. It plays the odds of increasing volatility ahead of earnings which you can combine with knowledge that some stocks more than others tend to move on earnings. Some other stocks are in fact immobile and these groups of stocks are not static forever either. Choosing the most likely candidates is therefore essential. I have done a couple of such trades - they have by and large either lost no or very little money and a few were good ones. You seem determined to bash the idea - its not a bad idea and a safer approach than many other option strategies. Its human nature to look for patterns in everything - even in chaos a human will discover patterns - so as far as that is concerned you are right to be skeptical. In this case though I would suggest to be a little less sarcastic and more open-minded.
How do you figure out the stocks are likely candidates for this strategy? Has there any academic literature regarding these. I looked at aforementioned Jeff Augen's book. It has one chapter about "Trading the earnings cycle" and in that he has 2 (two !) anecdotal examples of GOOGL and AAPL earnings. So here we have strategy which is so successful, but no literature. I am not bashing anything, but i am skeptical of any consistent outlier returns of 81% CAGR with SR over 2. Those are not characteristic of Long vol. returns. Short vol. will have high SR along with big draw down during times of stress. There has to be some intuition about any pattern whether structural as in case of overpriced index puts or behavioral like momentum. I also provided another pattern post-earnings drift, because of some behavioral reasons. There is about 5 to 6% per year before cost that can earned from these drifts. But if you include 2% month transaction cost along with any funding cost then those returns are difficult to earn. Same thing applies for long vol. before earnings. If indeed the vols are under-priced consistently on certain stocks 1 week before earnings, then this means market makers on those stocks are wrong consistently and even if they under-pricing do exist it is difficult to exploit after retail transaction and funding costs and security selection. I am not sarcastic about the strategy but definitely sarcastic about any outstanding returns.
I'm not defending anyone here and am not a client but I don't see how a lack of academic interest in a very specific trading strategy matters. Write your own white paper on it if you're so interested. Most options related papers are related to vol modeling and market structure, not something this specific and practical.
The way I traded it is using a back test against a number of companies with upcoming earnings. I made such trades as recently as September/October for AZO, ATVI, GOOGL - the google trade brought a small loss. The others profits - I am planning to do this again where I find the stocks that have momentum and who have shown movement ahead of earnings in the past. The idea is that the skew in the pool of choices will yield a higher than average number of stocks that will yield a profit on the straddle. The ones that dont move are hopefully protected by the increase in implied volatility. As regards Jeff Augen's book, I know it quite well. The examples of Google and Apple are not anecdotal - he says quite literally there are hundreds such stocks exhibiting such behaviour. Jeff is just realistic in his books in that the strategies he proposes have been fleshed out only for a few stocks - usually about 5. It takes a lot of number crunching to verify the spikes, crosses and other things needed to identify the right candidates and the right time for those candidates. I have replicated some of his work on spikes and crosses and can attest that you get quite tired after doing that for a handful of stocks regardless of modern technology. It also requires access to minute by minute quotes which are not available routinely for free. Jeff's approach is you are better off focussing on a few stocks whose behaviour you get to know better.
It's all about backtesting. You need to find out which stocks to use, when to enter and at what prices. For example, HD was been a consistent winner if entered 3-4 trading days before earnings. But we don't enter if the price is too high compared to previous cycles. Other stocks might show best time to enter as 7-10 days before earnings. Some of our members developed proprietary software that helps us to do the backtesting and see if the price is right. It helps us to identify the right candidates at right prices and right timing. For example, we entered CRM straddle few days ago based on the following chart: In some cases we might enter early and sell weekly OTM strangles at certain ratio to help to fight the theta. The strategy produces 5-7% average gain before commissions (commissions usually reduce the gain by ~1%) with very low risk and average holding period of 5-6 days. It is very consistent over time.
Ya, My paper trading of hindsight long vol. strategy had 100% hit rate as well with a SR of over 3. Institutional investors were interested and asked for verified account statements. I told them you have to go through captured screen shot trade confirmations, never heard back from them. I guess i have to settle for $125 monthly subscriptions from retail.
Regarding Jeff Augen's back, I will say this. If I am writing a book and present a hypothesis, I would atleast test that hypothesis and present the data or reference a study or paper which did that test. Not just presenting only 2 earning cycle vol expansion examples and hand waving and stating it may exist in other stocks. You can filter daily standard deviation spikes that you are referring from many softwares these days. OptionVue has it. Good luck with your trades.
Just tested entire nasdaq 100 stocks long straddle for last 3 years. Used CMLviz trademachine pro: Not sure about the quality of data. Criteria tested- Custom earnings: Goes long end of day 3 days prior to earnings, closes at the eod 1 day before earning. Buy and sells at market b/a. Weights: Equal weight for all the stocks ($10,000 worth of straddle) Results: Entire account of long straddle lost about 46.4%. So it is nothing to do with market structure. FAANG stocks were slight winner. Probably with good screening like quality earnings momentum, may be there is slight positive alpha. No where near great success story as it touted as.
LOL - that's the wrong criteria for starters. You have to look between 14 days to 7 days or so before earnings. Implied volatility ahead of earnings begins to trickle in at that moment. The biggest sweep up in the volatility occurs in the last 7 days so if you look at 3 days you are buying heavy premium. The second mistake you are making is believing this is some kind of magic always win for every stock strategy. Specifically stated was that you have to back test against specific stocks. As I wrote some stocks are in fact static ahead of earnings moving neither right nor left - you will lose money on those and lose big time. So doing a scan of the Nasdaq 100 is meaningless because you should have looked inside those 100 stocks and chosen the ones that have exhibited this behaviour in the past. Third mistake is standard closing 1 day before earnings - this has a major effect for stocks that report AFTER closing as opposed to before. What makes me laugh though is that you used CMLViz to do your testing.. They are the ones that routinely recommend this strategy in their regular updates. In fact they suggested you open a straddle on TGT yesterday in their news mail of end October.
This is kind of studies that tastytrade do. They are meaningless, for few reasons. CMLViz uses EOD prices. We use intraday limit orders for entries and exits. You cannot just take 100 stocks and test them. This is completely meaningless. We have a very specific list of stocks that we use cycle after cycle, but even for those stocks, we enter only when conditions are right. You cannot just select random days to enter end exit. For some stocks the time to open could be 3 days before earnings, for others it could be 10 days. The test completely ignores gamma gains. For example, if you start with 100 straddle, the stock moves to 105 and then goes back to 100 before earnings, you are missing very good gamma gains. I reality, we would book gains after the stock moved to 105, re-eneter the 105 straddle and book the gains again after it moved back to 100. Buy and sells at market b/a - this criteria alone probably reduced the gain by at least 3-4% per trade. In reality, the fills are much closer to the mid than to the market b/a. But then again, I don't need to prove anything - our results speak for themselves.