Billy,
this is very interesting. Is your position sizing independent by the strenght of the robot signal?
Billy,
this is very interesting. Is your position sizing independent by the strenght of the robot signal?
Billy,Originally Posted by Billy
This is a terrific explanation, and I will study it with intention to implement a similar approach.
In fact, I had worked out with Paul's help a similar though less sophisticated approach in late September and early October. In essence I placed 1/3 in each of the following: 3X, 2X, and IWM, such that the exposure equaled a 125% position in IWM, long or short. Any sudden price spike in the 3X ETF would be taken as windfall profit with the hope to re-enter the next day if a signal permitted. This allowed a suprising outperformance of the robot.
However, I sought to follow the robot more precisely in November and failed to grab the ready outsized profits on the occassions that appeared only briefly but obviously in the 3X ETFs.
The question of course becomes: Do these human inventions warrant backtesting? Your formula is extremely precise, and I believe mine can be made so-- although as noted previously I cannot translate my procedures and intutions into complex mathematical or logical operations.
Lastly, in answer to the above question about Dr. K's MDM: When I saw what Pascal, Billy and team were assembling here, I knew immediately it offered a golden opportunity to develop my skills. The styles of traders are varied, the discussions are rich, and the wizards are willing to help.
I did have one problem with the robot. I didn't realize it at the time. I had been closely and successfully following Morales and Kacher. While I grew tired of attempting to hop on break outs and ride them for profit and I wanted to learn a style that would give me more precise entires and exits, I had internalized their approach. Thus, when the robot issued a buy announcement and the price had not formed a cup or base in June , I couldn't compel myself to follow it. I kicked myself as I missed substantial opportunities. Indeed, I would say my choice to switch has been very costly in the short term, but I expect it to be a very good long-term decision.
Thanks all,
Last edited by nickola.pazderic; 11-24-2011 at 10:38 AM.
Hi Billy,
Thanks for the excellent explanation.
I am interested in potentially trying your approach, so I have tried to see if I could re-create the various aspects of the trade.
I am struggling with the first support cluster and related initial stop on the 2/3 position. I went back to the Forum Clusters 111115 file you published before the open on Nov 15, and I see the first support cluster on IWM as 70.36 to 72.63. Accordingly, the bottom of the first support cluster on IWM at 70.36 is 3.5% below the Nov 15 open of IWM (72.92).
Your above note seems to say that your TNA stop at 42.10 was +0.2% below first support cluster on TNA - this is only 5.4% below the TNA open at 44.49. I would have thought there would be more like an approximately 3X % difference between the IWM and TNA stops, i.e. the bottom of the first support cluster on TNA more like in the area of 3.5%*3 = 10.5% below the Nov 11 TNA open??
If I traded TNA based on IWM price triggers (which I can do in my IB account), than an IWM trigger just below 70.36 (the first support cluster bottom) would have resulted in a lower TNA price than 42.10, perhaps in the range of $40 and exposed such a trade to about twice the downside risk you are describing.
What am I missing?
Thanks,
Shawn
Roberto, the various “strength” levels of the signals have been abandoned with recent developments of the robot. So all signals – buy or short – are now strong signals and can be approached similarly for position sizing.
Billy
Shawn,
You are correct. There was a typo in my journal notes and the initial stop was 40.10. Thank you for noticing and for the feedback.
Billy
Nickola,
I want to stress that what I am presenting here is just the approach I feel most comfortable with in the current choppy environment. There are numerous leverage and position-sizing strategies that I use at different times and most are very discretionary with a lot of tolerance for porosity at key levels. If you remember, before the robot issued that last long signal and setup, I was already very bearish personally. I was doubtful of a successful trend-following trade. Since I’ve rarely seen a long signal of the robot not being in the green at least for one day or two, I was determined to grab a good chunk of the leveraged profit if it retraced more than 50% of an initial progress. The definition of an uptrend is higher lows and higher highs. The definition of a sound uptrend is one that doesn’t retrace more than 50%. Hence my choice of the current leverage management strategy.
My goal is always to at least equal the performance of a non-leveraged IWM position on average over time. The first condition I know to do better consistently with triple leverage ETFs is cutting your initial losses much faster if the trade turns wrong just after entry, but allowing for reasonable wiggle room. One support cluster sounds like a logical tolerance for initial stop, since it encompasses all support levels that could be hit within one day. The second condition is to book most of the gains compounded with a 300% factor before they reverse to losses compounding with a 300% factor.
Creativity and the list of possibly valid strategies are endless. In the end It’s all a matter of feeling comfortable with your logical plan and you need to adjust the rules with market conditions. In this leverage field, experience is more important than backtesting, because backtesting will have a hard time finding adaptive rules for changing trading environments.
Enjoy your stuffed turkeys everybody!
Billy
It is very interesting to me that Paul and I worked with a similar strategy in late September and early October. I realized a 17.5% increase in the TOS account in a matter of days. I mentioned this outperformance to you in private. Either I did not communicate precisely my meaning or you misunderstood it as a conscious attempt to outperform the robot or both. Actually, I was doing nothing other than attempt to ensure that my equity curve remained in an uptrend. If I can realize an equity curve rise like that displayed on our home page for IWM, of course I'd be delighted.
Because our communication had a flaw, when the November trades came, in particular the one you mention, I recall being torn between taking the profit and following diligently. I did a little of both and lost a lot more than planned.
The Fibonacci idea seems like a sensible way to draw stop losses on a exploding 3X instrument, and placing variable stops that ensure against blowouts to the downside is also most a helpful idea.
This discussion has been very helpful for me. I won't fear the robot signals nearly as much as I did during the past two trades when it felt like the The Charge of the Light Brigade as we went long into the jaws of a bear.
I'm a vegetarian, but I'll find some things to eat as I drink Argentinian Malbec with relatives.
Best to you,
Last edited by nickola.pazderic; 11-24-2011 at 02:04 PM.
Yes, I was thinking to the different ST and LT edges of a confirmed signal.
I then suppose you treat them all the same.
Posting Permissions
Reply With Quote