Using options for hedging purposes

[trefoil]

This is a time frame when the Nasdaq spent the majority of its time headed south bigtime. That would mean the test results would indeed show the tighter the collar the more you make. Kind of elementary.
A better test would be a time frame like Apr 99 to Mar 2001, which would have given you a more or less equal amount of time on either side of the Nasdaq peak in Mar 2000.
This is indeed a cherry picked time frame.

 

[Tomaz26]

Quote from trefoil:

This is a time frame when the Nasdaq spent the majority of its time headed south bigtime. That would mean the test results would indeed show the tighter the collar the more you make. Kind of elementary.
A better test would be a time frame like Apr 99 to Mar 2001, which would have given you a more or less equal amount of time on either side of the Nasdaq peak in Mar 2000.
This is indeed a cherry picked time frame.

This is not true. Please see the picture. They picked march 1999 because this is the first time options on QQQ were available. They`ve chosen 2002 because it is the lowest point after the big sell-off. This period 1999-2002 also has a very big run-up included, I guess this is where the majority of return comes. So the point of this period was to show that with collar you can participate in bubble run-up and protect some or most of the profits after the bubble burst. I think this could be the strategy to use for gold lately. No one knows when this run-up will end and if you want to be protected if the big sell-off comes..
Please see the picture.

They also did an out of sample test on later period/s and also out of the sample test on mutual fund..

 

[spindr0]

Quote from trefoil:

This is a time frame when the Nasdaq spent the majority of its time headed south bigtime. That would mean the test results would indeed show the tighter the collar the more you make. Kind of elementary.

I don't think so because with a collar, you lose on the UL down to the strike. Below that you break even.

If you were to do make believe collars where the UL was always at a strike and that strike was your long put leg, the option collar would always be put on for a debit and the same result as above, a loser as the UL drops.

No passive collar (standard bullish vertical) makes 23% per annum in a collapsing market w/o other factors such as extra puts, overwriting the calls, etc.

 

[joneog]

Quote from Tomaz26:

There there. Now that you say it this way, I am convinced that this research paper is just a load of crap. Can`t argue with your strong arguments and facts.. Surely some on-line forum guru will know better than a bunch of guys with their fancy PHDs..

Here's the order you should follow for investment/trading advice:

monkeys throwing darts at the WSJ > internet gurus > CNBC gurus > infomercial gurus > finance/econ PHDs

And it's still a little insulting to the monkeys.

 

[atticus]

Quote from Tomaz26:

There there. Now that you say it this way, I am convinced that this research paper is just a load of crap. Can`t argue with your strong arguments and facts.. Surely some on-line forum guru will know better than a bunch of guys with their fancy PHDs..

ESL Genius,

I received my B.Sc. from Chicago at 19... while you were finishing your second shot at senior year of HS.

The argument hasn't changed. You cannot manufacture a 23% return on a bull vertical spread that is always in the market during the period in question. Either it's completely fabricated or they cherry-picked the backtest by converging (strikes) to a synthetic short and/or going to cash. A massive hindsight bias.

Apparently the PhD wasn't enough. They could've saved 30 pages by backtesting the vertical in lieu of the collar. Best practice it ain't.

 

[spindr0]

Quote from Tomaz26:

So the point of this period was to show that with collar you can participate in bubble run-up and protect some or most of the profits after the bubble burst. I think this could be the strategy to use for gold lately. No one knows when this run-up will end and if you want to be protected if the big sell-off comes..
Please see the picture.

They also did an out of sample test on later period/s and also out of the sample test on mutual fund..

You could test this easiuly with historical data. Since that's usually expensivve to come by, you could do some hypotheticals in a spreadsheet. Easier to test a hypothetical prices, more effort to test with actual UL expiration values (stock is known, would have to determine option prices with static vol). I'll save you the time. Not going to get 23%

As for the 50 page paper and the stats sheet, I'm crosseyed enough from staring at platform prices all morning

 

[Tomaz26]

Quote from spindr0:

I don't think so because with a collar, you lose on the UL down to the strike. Below that you break even.

If you were to do make believe collars where the UL was always at a strike and that strike was your long put leg, the option collar would always be put on for a debit and the same result as above, a loser as the UL drops.

No passive collar (standard bullish vertical) makes 23% per annum in a collapsing market w/o other factors such as extra puts, overwriting the calls, etc.

You are correct, this is the time when QQQ first rallied 100 % and then collapsed 81 % so in this time you made 23 %, which if I understand could be possible. I attached the picture in my previous post where it can be seen how QQQ performed during this time. Like I said they picked this subtime frame to show you can participate in a big run-up and protect profits in a collapse that follow. Ofcourse if you would use this strategy at the top without participating in run-up the picture would be different, but that would then really be cherry picking. They just wanted to show this strategy is usable in 2 scenarios, first when there is first a big run-up and then brutal sell-off(99-2002) and second reverse story when there is first a big sell-off and then recovery(2007-2010). The strategy performs poor in their third subperiod 2002-2007, which they clearly show. This is the time of steady growth, low volatility and low almost no major sell-offs. It is normal that in that case you are best off just beeing long UL, even better of long on margin... But since I dont believe those times are anywhere near the corner the stategy could be usefull to implement now, specialy for those wanting to protect profits from big run-up...

Also they did the test for the whole period, from the start of the options on QQQ to the 2008 and now extended to 2010. So they did not cherry pick and subperiod to show how good their stragegy is, they just showed 3 subperiods so readers can see how this strategy performs in different environments. All in all if you use this strategy long enough, through bull and bear market cycles, you should get much better returns and much lower volatility then buy and hold. The idea I also like is that it does not involve market timing and trying to pick tops and bottoms. Just hedge with collar all the time and except that you will trail in the bull market but come out ahead a lot in the bear. One should sleep a lot easier like that...

 

[atticus]

Quote from Tomaz26:

I found a very good paper testing different collars for QQQ from 1999 to 2010 with very good results. This strategy of course trails returns in bull market(returning 5-7 % yearly instead of 21 %) but really shines in nasdaq decline in 2000 and also very good returns from 2007-2010 ! For example in declining years you get +23 %.

 

[Tomaz26]

Quote from atticus:

Yes, my mistake, I wrote declining years as a part of this 99-2002 subperiod, because I thought everyone allready knows that Nasdaq did not just decline from 99-2002 and that it also had a spectacular run-up. The part I wanted to stress was a decline and that you can also protect your profits even if there is a period of a big sell-off. Ofcourse that also meant that you still have to participate in some of the run-up.

As far as I can see, the strategy performs poorly only in long bull markets with low volatility, for other times it performs very good

 

[trefoil]

Quote from spindr0:

I don't think so because with a collar, you lose on the UL down to the strike. Below that you break even.

If you were to do make believe collars where the UL was always at a strike and that strike was your long put leg, the option collar would always be put on for a debit and the same result as above, a loser as the UL drops.

No passive collar (standard bullish vertical) makes 23% per annum in a collapsing market w/o other factors such as extra puts, overwriting the calls, etc.

You are correct, I made a mistake there. But the results - relative to each other - are consistent in that as you move the call side from ATM to higher than that, the returns decline, which is what you would expect in a time period when the predominant direction was down.
I don't know how you would make that 23% though. You could be positive if you get a bounce after selling the put and then sell the UL. But that would be on that one trade. Doing that consistently for three years would be tough.