Discussion in 'General Discussion' started by Ice101781, Apr 1, 2017.
Time for a large-scale automated back test! Can you provide the precise parameters of the trade?
Sure. Quote from his book "We have tracked the performance of a short 1x2 put ratio on the SP500 over a 10 year window. Specifically, we sold 1 25 delta put and bought 2 10 delta puts, resetting the structure weekly. Both strikes had 4 weeks to maturity at initiation. Our results are summarized in Figure 4.9"
I have to correct the sentence "delta neutral". It is not delta neutral in the above graph. His favorite structure is 2x5 (2x25delta short, 5x10 delta long) which is delta neutral. I got mixed up in many 1x2s he tested. Similar graph for FTSE.
Ok so it's a long delta trade - that totally changes the picture since simply being long deltas is a great strategy over long time periods. Does he provide long-term results for the 2x5?
He has several interesting graph's in his book. Here is another regarding 1x2 structure.
Steve. Unfortunately no.
I initially disagreed with your post in the following area.
He has graph after adjusting for in-sample volatility 10 deltas puts are cheapest compared to near the money or 25 delta puts, so that is better suited for hedging. Also you see in the above charts effective hedging is not very expensive (3.1% per year)
There were few 5 to 10% gap down during the testing period. 1x2s seems to have performed better
Thank you to whoever recommended the book mentioned above. I find it excellent and am actively testing strategies he discusses in the book. I have already been using call backspreads in the vol products for quite some time with good success. By good success I mean they have allowed me to profit from my short vol side, not that I have made money on the back spreads themselves yet
I have no disagreements with anything you are saying as I have not done sufficient back testing yet ... my main angle with back ratios is to construct the hedge you really want in terms of (A) decreasing max loss and (B) zeroing out the catastrophic risk without buying more downside than you need ... then see how much this hedge costs and what, if anything, to do about the cost. If your hedge costs you one-third of expected profits on average but it means you can double your trade size with lower net risk, it makes some kind of sense.
I do think it's a rule with back ratios that going too far in strike width is toxic ...
For what it's worth, here's a wide-angle snapshot of a back test of 2x4 SPX back ratio with upper strike at various deltas, entered 55 DTE, 30 DIT, strike width 50 points, dynamically hedged, and rolled whenever the market moves 1%. The p&l numbers are in SPX points. A total of 301,712 trades are represented:
Yes, although he talks about 25 delta and 10 delta I find at least so far in my analysis that keeping the width to 50 points is advisable, wider is as you say "toxic"
And let's not forget Ron's 3x5, he thinks that is a better structure so I'm testing that variant as well
But Ron's trade is long DTE, far OTM and not designed to effectively hedge atm structures ... I'm not sure he has strong opinions on how good (or not) a shorter-DTE, nearer-the-money version is. For what it's worth, here is the same back test pictured above (2x4s, blue) with the same thing on the right except 3x5s (yellow). I look at the 3x5 as a 2x4 with layered-on credit spread ... you give up some of the nearer-strike hedging power to get a cheaper hedge. I also view the 2x4 versus 3x5 fluidly: buy back the credit spread if you're nervous or need a few negative deltas; sell it if you're complacent or need some positive deltas.
Yes, I like the way you look at the structure peeling it into it's layers, I assume you've read Cottle, Saliba etc.
I would tend to agree with you that he is too far out in time, but he has convinced me that it may have merit so I'm looking at it as well. I use call back spreads in VIX also, I think the book gets into those as well but I've only read about the first half or so, just got it two nights ago
First, thanks to Ice101781 for opening up the thread. Great discussion.
Also thanks to Srini for mentioning Hari Krishnan's book. I've been alternating between reading through the kindle edition and running a few spot check back-tests on how the put-ratio performs in a few different environments.
To Krishnan's credit he is very easy to read, and I like that he takes the time to point out the practicalities of the approach, for example:
I also like that he has actually used it in practice:
The short 1×2 put ratio is something that has been in our arsenal for quite some time. It worked particularly well during the Flash Crash of May 2010, when the put skew steepened dramatically for equity indices and risky currencies.
Having said that, a close reading of one of the book reviewers, shows the alternative view....position sizing:
Praise for The Second Leg Down
"The most difficult portfolio management decisions come after an unexpectedly large loss in an unsettled market. This book tackles that situation with a comprehensive, accurate and clear account of clever strategies to navigate the troubled waters with little or no sacrifice of upside. It's well-written and entertaining. Most of the time you're better off with the unclever idea of cutting positions to where you're comfortable for the risk, but if you refuse to do that, this book gives you the best chance of profiting from your aggressiveness."
—Aaron Brown, Chief Risk Officer, AQR Capital Management
Yes. Thanks to the OP for posting the topic.
Since i carry futures position overnight, I am always worried about sudden reversals. Being long equities, short bonds and risk currencies is worst possible scenario, when s@#t hits the fan overnight. Correlations break down during that time. One may not have time to cutting the positions as Aaron Brown says.
Also most of AQR runs their funds at risk target of 10 or lower. Even unlevered SPY has more risk than their funds. It is very easy to shed positions at such low risk target. AT elite trader "globalarbtrader" runs his account at 25 risk target. I would say most of the retail traders is in that range.
To be fair, 1x2's is been their long time and used by many. But i was impressed array in which he employs. Along with 1x2 put spreads, 1x2 in call spreads in bonds and short future VIX+Vix calls. I have done in short VIX puts+VIX call spreads before. Short VIX future+VIX calls is much better strategy. Employing all these together is much better suited for me, since correlation breakdown does not effect having all 1x2 equities, bonds, vix together.
Also I was never much big fan of weekly options. Now his saying of "Vega before, Gamma after" is ingrained in me to just take a fresh look at weekly options.
Reading this thread really underscores for me how little I know about long term hedging, how much I'm exposed to a major move down, and how far I have to go just to understand this discussion.
Thanks for the book recommendation and I'll try to understand the 1x2 and strategies discussed here better. Not knowing what I don't know is dangerous, but at least now I know what I don't know...
A very good point....overnight gaps are the enemy of an optimized fixed fractional or even half-kelly money management system.
It's an awkward point, but I do believe there is gravitational pull once a trader starts expecting much beyond 15% net compounded annual growth rates, as discussed here:
There is no evidence suggesting you can obtain higher than 30% returns year after year for decades on end. In fact, only very few cases throughout history have been able to achieve this at the professional level. And not even the almighty Buffet.
The Barclay's Discretionary Traders Index is averaging 7.37%/year after three decades. https://www.barclayhedge.com/research/indices/cta/sub/discret.html (with some survivor-ship bias, as funds that under-perform and die stop being tracked).
What we do know, is that 90-95% of traders lose money, and that anywhere from 75%-90% of professional money managers fail to beat the S&P500 on any given year. This means that a +10% avg annual return in the long run puts you at an elite level, as you make money (therefore beating 90-95% of traders), and you beat the S&P500 (beating the vast majority of professionals). +15% is great. Anything above 20% is awesome. Its the cold, hard truth that those who know better, don't want to talk about.
This is a very sensible and realistic view of what investing and speculation can yield in the long term. Way back in 1993, Lawrence Harris had published a paper explaining why most market participants must lose money in order to keep the markets going.
These guys https://www.poweropt.com/ourcompany.asp promote and advocate some fairly realistic and achievable strategies. They wrote a book about it. https://www.amazon.com/Protective-Options-Strategies-Married-Spreads/dp/1592803423 is conservative and prudent though not very exciting.
Just a FYI. If I am not mistaken, I think Ron's 3x5 is a long (debit) ratio spread, not a short (credit) ratio spread or square root spread illustrated in the book.
I maybe misinterpreting. I have a suspicion that you interpreted "risk target" as "return target". My apologies if it is not the case.
To clarify for "risk target", it is nothing but "volatility target". Most of hedge funds (at least good ones) and risk parity strategy (mentioned and warned in the book) invest based on risk allocation, not fixed dollar allocation. Measurements of portfolio allocations by dollars could be highly misleading. Traditional 60-40 (stocks & bonds) sounds balanced, but has 90-10 risk!.
Risk parity strategy achieves diversification, by leveraging lower vol. assets (in this case bonds) to bring it to same level volatility as equities. This is based on historical performance and correlation. What if volatility and performance breaks down after 35 year rate drift down? To control this they target certain volatility levels. AQR targets volatility of 10. This means when ever realized volatility of the individual asset of past certain days exceed 10, they bring down the weights of the asset so that volatility will be equal to 10.
Historically these strategies have done well. One because of bonds had terrific performances with low vol. Second because equities position was significantly reduced just prior to black swan events. If you go back and check, realized vol doubled just before 1987 crash. That means equities allocation would have been halved.
Here good links regarding this.
In fact, I would read anything put out by AQR.
Apologies accepted....thanks for the awesome AQR links...most appreciated.
After reading this book: http://qoppac.blogspot.co.nz/p/systematic-trading-book.html my eyes were opened to a lot of volatility targeting approaches and implications.
You might already have Rob Carver's book.....I was going to post a few scans from it, but I'm not sure how UK copyright laws work
He was running AHL's systematic bond fund back in 2008....a really interesting read (at one point they had a $1B up day in Sept 2008...freaky stuff!!)
One quote from his book about systematic trading won't hurt though:
"....be nervous. Only commit trading capital you can afford to lose. use Half-Kelly: your maximum percentage volatility target should be half what you pessimistically expect your Sharpe ratio to be. At some point some instruments or trading rules in your portfolio will go badly wrong."
His blog is really interesting if anyone wants to dive into lots and lots of content, in particular his recent QuantCon presentation (all 52 slides of it )... http://qoppac.blogspot.co.nz/2017/05/some-reflections-on-quantcon-2017.html
Separate names with a comma.