Watch HowardCohodas Trade Index Options Credit Spreads

[HowardCohodas]

Howard old chap, you say you'll give up after 3 events... over an infinite time that will almost surely happen. So this strategy is flawed, right?

There is a small but finite probability of a lot of things happening, like tossing heads on a fair penny 1000 times in a row. I don't think we are arguing the impossibility of the event. We are trying to discern it's likelihood. If the probabilities are such that it wont happen in my lifetime, then...

The point of mentioning even unlikely events is for completeness. But then some effort ought to be given to the probabilities. I seem to be always told that I have not listed all the outcomes, when I believe I have. We seldom get to the useful part of the discussion.

 

[arabianights]

There is a small but finite probability of a lot of things happening, like tossing heads on a fair penny 1000 times in a row. I don't think we are arguing the impossibility of the event. We are trying to discern it's likelihood. If the probabilities are such that it wont happen in my lifetime, then...

Then what? Once you accept the possibility of blowing up you can make any strategy profitable.

 

[the hare]

We are trying to discern it's likelihood. If the probabilities are such that it wont happen in my lifetime, then...

Jesus Christ !

Assume there's a probability of a feckin huge storm occurring once every 10,000 years.

That doesnt mean the storm will come along and blow your house down 10,000 years from now. The storm could strike tommorrow. A similar sized storm could hit again a month after, and not strike again for 100,000 years.

FFS buy a book on basic statistics... then feckin READ IT !

If there's such a thing as Karma you, and every other vendor who abuses basic statistics to fool the gullible deserve to drown paddling in a lake whose average depth is 6 inches !

 

[amit1986]

Jesus Christ !

Assume there's a probability of a feckin huge storm occurring once every 10,000 years.

That doesnt mean the storm will come along and blow your house down 10,000 years from now. The storm could strike tommorrow. A similar sized storm could hit again a month after, and not strike again for 100,000 years.

FFS buy a book on basic statistics... then feckin READ IT !

If there's such a thing as Karma you, and every other vendor who abuses basic statistics to fool the gullible deserve to drown paddling in a lake whose average depth is 6 inches !

Buddy, take a chill pill - we're all just trying to have a civil conversation

As for the topic in discussion, I would think Howard understands the risks associated with his strategy quantitatively as he was an engineer.

There is a very small chance of a blow-up happening, but if it is a very low probability, then why not trade it? Aren't we all in this game to find a high probability strategy? There are no strategies that will give a 100% success rate...

Howard cannot change the probabilities but he can be smart about his risk management. One way to do this is to make sure you set aside some profits from your account (a fixed percentage every month) incase a blowup does occur and it would leave him out in the cold...but I am sure Howard already does this...

 

[HowardCohodas]

Yes, HC, unfortunately I can't put numbers arnd my assertion (yet).

The most important thing isn't VIX per se. What's important is in the attached graph.

As to the sensitivity analysis, it's rather meaningless, innit? In a particular regime we're in, I can imagine it doesn't really matter whether your parameters are tweaked a bit. As to the out-of-sample testing, that's actually quite interesting. Given you have done this, are you then suggesting that your specific rule/parameter combination will work in any chosen mkt regime? Finally, what do you mean by a rule/parameter candidate? Are there several? How do choose one over another?

The reason for so many iterations over the back-test data set was to do the sensitivity analysis over a set that covered a wide variety of market conditions. I think doing a sensitivity analysis when qualifying a strategy is as essential as validating it with a subset that the strategy has never seen. Otherwise, I would not have any reasonable confidence that I have something worth the next step in the qualification process, i.e. paper trading.

When I did the back-testing, I did not examine the results at the level of the market conditions for each spread. I looked at the usual suspects in metrics to judge the "goodness" of the results.

Because back-testing and paper trading have so many limitations, I'm not prepared to state that the strategy as it is today will survive all market regimes. Only real money trading will give that level of confidence.

Regarding my starting point for the parameters, that was primarily based on some research and some instinct.

 

[HowardCohodas]

Once you accept the possibility of blowing up you can make any strategy profitable.

 

[j b]

Even if a full account blow-up is highly unlikely, that doesn't stop the occassional losses being large enough to stop the strategy being profitable in the long run. There still doesn't seem to be any explanation of why this should be profitable, unless he's taking the view that options are perpetually overpriced.

As for the topic in discussion, I would think Howard understands the risks associated with his strategy quantitatively as he was an engineer.

Engineer or not, he doesn't understand the basics of option theory, so why would he understand the risks?

 

[HowardCohodas]

Jesus Christ !

Assume there's a probability of a feckin huge storm occurring once every 10,000 years.

That doesnt mean the storm will come along and blow your house down 10,000 years from now. The storm could strike tommorrow. A similar sized storm could hit again a month after, and not strike again for 100,000 years.

FFS buy a book on basic statistics... then feckin READ IT !

If there's such a thing as Karma you, and every other vendor who abuses basic statistics to fool the gullible deserve to drown paddling in a lake whose average depth is 6 inches !

From your comments, I'm not convinced I am the one who needs the lessons on probability.

For example, there is a small, but finite probability that all the air in the room you are in will move to the ceiling corner closest to due North. However the calculation shows that it's likelihood of happening even once is longer than the know age of the Universe.

So are you going to base any decisions in your life based on the this lesson in probability. I would doubt it. I know, you will now only stay in round rooms.

 

[Martinghoul]

The reason for so many iterations over the back-test data set was to do the sensitivity analysis over a set that covered a wide variety of market conditions. I think doing a sensitivity analysis when qualifying a strategy is as essential as validating it with a subset that the strategy has never seen. Otherwise, I would not have any reasonable confidence that I have something worth the next step in the qualification process, i.e. paper trading.

When I did the back-testing, I did not examine the results at the level of the market conditions for each spread. I looked at the usual suspects in metrics to judge the "goodness" of the results.

Because back-testing and paper trading have so many limitations, I'm not prepared to state that the strategy as it is today will survive all market regimes. Only real money trading will give that level of confidence.

Regarding my starting point for the parameters, that was primarily based on some research and some instinct.

HC, I'm sorry, but this is all neither here nor there...

Firstly, you haven't actually shown any results of your backtests here, so I have no idea if it's well-specified or meaningful. Secondly, this is not about the backtest being able to demonstrate that the strategy is viable under all mkt regimes. As you correctly point out, backtests and paper trading don't offer confidence that the strategy is 100% robust. However, that's not the point. Point is that the backtest should have demonstrated to you that a) the strategy breaks in certain mkt environments; and/or, b) the strategy isn't stable wrt the parameter values.

So in light of the above, let me try to help you. I'm gonna try to get smth together, but pls bear with me, as it might take a wee bit of time.

 

[HowardCohodas]

Even if a full account blow-up is highly unlikely, that doesn't stop the occassional losses being large enough to stop the strategy being profitable in the long run. There still doesn't seem to be any explanation of why this should be profitable, unless he's taking the view that options are perpetually overpriced.

Engineer or not, he doesn't understand the basics of option theory, so why would he understand the risks?

I guess the summary of results from nearly six months of trading this strategy has no impact on your thinking. How do you account for my results? Or are you a member of the class of people who operate by "My mind is made up so don't confuse me by the facts?"

 

[HowardCohodas]

HC, I'm sorry, but this is all neither here nor there...

Firstly, you haven't actually shown any results of your backtests here, so I have no idea if it's well-specified or meaningful. Secondly, this is not about the backtest being able to demonstrate that the strategy is viable under all mkt regimes. As you correctly point out, backtests and paper trading don't offer confidence that the strategy is 100% robust. However, that's not the point. Point is that the backtest should have demonstrated to you that a) the strategy breaks in certain mkt environments; and/or, b) the strategy isn't stable wrt the parameter values.

So in light of the above, let me try to help you. I'm gonna try to get smth together, but pls bear with me, as it might take a wee bit of time.

I would not have been that detailed about back-testing except to answer question that you asked in the post to which I was responding. I actually agree that the qualification process does not really answer some key questions you and I have already outlined?

Thank you for helping us all try to understand what is going on.

 

[Martinghoul]

I guess the summary of results from nearly six months of trading this strategy has no impact on your thinking. How do you account for my results? Or are you a member of the class of people who operate by "My mind is made up so don't confuse me by the facts?"

Howard, the fact that your strategy has "worked" for six months is a total red herring. One of my colleagues made lots of money every year from arnd 2002 to 2007. That's results from nearly five years of trading. He blew up in 2007 and lost more money in a couple of months than he made in five years. He was trading mainly options. How do you account for that? It's a well-known fact (and you should be familiar with it) that short gamma strategies often have skewed payout profiles (like selling insurance). So six months of good performance is totally irrelevant and you should stop mentioning this fact altogether, 'cause it makes your whole method look silly.

I would not have been that detailed about back-testing except to answer question that you asked in the post to which I was responding. I actually agree that the qualification process does not really answer some key questions you and I have already outlined?

Thank you for helping us all try to understand what is going on.

Well, I view it as mostly helping myself. Again, let me try to demonstrate what I mean, since it's probably better than making unsubstantiated stetements.

 

[HowardCohodas]

Howard, the fact that your strategy has "worked" for six months is a total red herring. One of my colleagues made lots of money every year from arnd 2002 to 2007. That's results from nearly five years of trading. He blew up in 2007 and lost more money in couple of months than he made in five years. He was trading mainly options. How do you account for that? It's a well-known fact (and you should be familiar with it) that short gamma strategies often have skewed payout profiles (like selling insurance). So six months of good performance is totally irrelevant and you should stop mentioning this fact altogether, 'cause it makes your whole method look silly.

Sorry to disagree with you. I think in context my response was appropriate. We agree, the results do not prove the strategy works. They serve as counterpoint to blanket statements that it can't.

And they are facts that must be explained. I don't know about your friend's experience and whether he changed his methods or money management or both. As you are want to say, it's neither here nor there when it comes to understanding what is going on or refuting categorical statements that it can't possibly work.

One recurring comment you make is regarding being short gamma. What order of magnitudes are we taking about? I'm confused on what levels should cause me concern.

 

[tar]

HC grilled on T2W ...

 

[meanreversion]

This thread has degenerated into bear baiting, lulz free zone now.

 

[j b]

Sorry to disagree with you. I think in context my response was appropriate. We agree, the results do not prove the strategy works. They serve as counterpoint to blanket statements that it can't.

But they don't serve as a counterpoint in any way whatsoever. If your selling vol, you're always going to have long periods of good performance - even if your long term expectancy is below zero.

The fact that a losing strategy can easily show similar performance to yours would suggest that your PNL is irrelevant - it shows what a good run looks like, but you're missing the other half of the picture.

 

[HowardCohodas]

But they don't serve as a counterpoint in any way whatsoever. If your selling vol, you're always going to have long periods of good performance - even if your long term expectancy is below zero.

The fact that a losing strategy can easily show similar performance to yours would suggest that your PNL is irrelevant - it shows what a good run looks like, but you're missing the other half of the picture.

I expose all my trades. Can you make a case for your statement or are you just generalizing. Without specific references to specific trades and the specific risks that I am taking that I have not already listed and without a history of establishing your credibility on this forum, one can hardly take your statements with much seriousness. They may be accurate, but what have you done to give them credibility? If you don't like my standards to establish your credibility, what standards would you suggest?

 

[arabianights]

Because back-testing and paper trading have so many limitations, I'm not prepared to state that the strategy as it is today will survive all market regimes. Only real money trading will give that level of confidence.

And how do you propose to test that then?

 

[arabianights]

Do you disbelieve me? Consider martingale on roulette... very easy to guarantee either profits or blowing up

Incidentally Howard there's been a noticeable drop in civility from yourself over the course of this thread... is that reaction to the way you're being treated or your true self coming out?

 

[Martinghoul]

Sorry to disagree with you. I think in context my response was appropriate. We agree, the results do not prove the strategy works. They serve as counterpoint to blanket statements that it can't.

And they are facts that must be explained. I don't know about your friend's experience and whether he changed his methods or money management or both. As you are want to say, it's neither here nor there when it comes to understanding what is going on or refuting categorical statements that it can't possibly work.

One recurring comment you make is regarding being short gamma. What order of magnitudes are we taking about? I'm confused on what levels should cause me concern.

I made no blanket statements whatsoever. I simply want to point out to you that, given the type of trading you do, your performance over a 6 month horizon should be the last argument in favour of your method. Preferably you should avoid an argument like this altogether, IMHO.

Pls note that I, for one, while skeptical, never stated with any sort of finality that your approach doesn't work. A priori, there's always a probability that you have discovered edge, but, with all due respect, it's rather small. Pls don't take offense, but it's akin to a first grader claiming that he has proved Fermat's Last Theorem. It's theoretically possible, but highly unlikely. This is not to say that you're stupid and all the skeptics here are smart, but a simple matter of odds and incentives.

As to short gamma, I use the term loosely to describe a whole variety of strategies that all, at their core, involve selling optionality. These types of strategies (not all of which necessarily entail explicitly selling options) are extremely difficult, if not impossible, to get right and history of finance is full of stories of people who tried and failed most spectacularly. Personally I have closely observed several such examples myself. What you're doing is a somewhat tamer version of such a strategy, but it doesn't make it any safer.