Algorithmic trading for hedge funds: hedging techniques and application to a folio

Discussion in 'Journals' started by fullautotrading, Jul 21, 2014.

  1. Dr Who

    Dr Who

    I really admire your tenacity but I'd give up and move on to something else. Like gardening for instance
     
    #91     Oct 30, 2014
  2. I will wait your expert thread on that, to learn from you. I do have a large garden which may benefit from that. Thanks.

    In the meantime, feel free to continue trolling and also maybe post some of your "curve-fitted delusions" and interpolations of the past.
     
    Last edited: Oct 30, 2014
    #92     Oct 30, 2014
  3. You certainly remember (see one of the previous post) I had placed a "trigger" ("enqueued order") on DGAZ to start trading only if the price went above a certain price. We said that we were willing to start trading it only if the price went "high" enough, due to the correlation with UNG, and the absence of options.

    Actually, that trigger price was touched these days, and it even managed to quickly scalp an easy profit, in between the large number of rejected orders of ERY and the others. I have now shut it down again because I cannot obviously afford it at the moment and need to focus all the activity to recover ERY (and, if possible, some of TNA and TZA). Anyway, I have placed again a new trigger, just in case:

    DGAZ_1.png

    This quick incursion suggests how it is possible to greatly reduce the risk profile (in statistical terms), if one had the patience to wait some decent moves (e.g., 40%+) before starting trading some instruments. Of course it's all matter of available capital: the more we have, the more we can afford to be "impatient" and start trading from relatively risky levels.

    Also, for the purpose of a thread illustration, waiting too long periods (long moves against the leveraged ETF drift) before starting trading an instrument may be not much entertaining. Anyway, for real trading with relatively scarce resources, it can be advisable, imho.
     
    Last edited: Oct 30, 2014
    #93     Oct 30, 2014
  4. can you explain in a short paragraph why your strategy should work?
     
    #94     Oct 30, 2014
  5. Hi trend2009,

    > should work

    What we can guarantee is the approach to make sense from the viewpoint of an expert manager (I am excluding clueless newbies, still living in their world of curve-fitted fantasies and of imaginary predictive "signals"). "Working" is another matter, which depends on several factors, including instruments, experience, available capital, psychological capacity to sustain the (inevitable) phases of DD, margin method, entry sizes, entry spacing, and the shape of the price curves that will take shape in this reality (which is an accidental component on which we have limited control). Obviously, there is always a risk profile involved.

    Nobody in the world can "guarantee" you will become rich, apart scammers or clueless idiots who never real traded. What you can (hope to) guarantee is to execute the most reasonable actions ("game rules") under the given circumstances, and not rely on delusional ideas or a "magic thinking" approach.

    > in a short paragraph

    Coming to your question, what we have been doing in this specific thread is essentially to take a long term position along the structural drift of leveraged ETFs (to be slowly unloaded in the future), while simultaneously - in the short term - scalping against the long term drift (hedging orders), to make more sustainable to hold on the long term position.

    [The novel part is the fact that we actually use past trading information to recover and "recycle" some of the losses due to hedging orders (thus replacing the concept of stop loss, with a more useful idea, where we physically store this information to inform the future strategic actions).]

    Hope it's concise enough: let me know if also clear enough.
     
    Last edited: Oct 30, 2014
    #95     Oct 30, 2014
  6. Clearly, using the past trading information is the crucial part of the method. In fact, even it it's "just information", it is crucial and important information to be able to mechanically recover the past losses.

    Without the quantitative information stored in the "players" one would just know he accumulated "some" loss, but would be completely lacking the quantitative information for a precise and scientific recover, just where it is needed.

    Let's take for instance ERY, which, as we know, had a good DD and look at the situation. The picture shows the open buy (cyan) and sell (red) players. In this case we retain the precise information about the loss (in this case mostly caused by the BUY players which could not be closed due to order rejection by the "too strict" margin method). Precisely where it happened, how much it is singularly and globally for the side (buy side in this case, and about 460K) and therefore we can possibly "unwind" it "point by point" (order by order) in the future (if the prices goes back there after closing the sell players), or (in any case, even if the price never gets there again, use it to appropriately resize the next sells). [The loss "unwinding" is what, as previously discussed, feeds the positive PNL drift. Clearly, with structurally drifting instruments (like those leveraged ETFs), there may be less losing players which can be "resolved", because the price may never "get back" there (and slide through "splits" or reverse splits), however they generally provide faster price waves which are more suitable for scalping.]

    If you compare this with the ordinary "stop and forget" approach, you readily see how the manager, even obviously conscious he has accumulated "some loss", would be mostly incapable to precisely "recover", simply because he misses the precise quantitative information. Nor he has a precise and automated mechanism to integrate the knowledge of past trading information in the future actions.

    Clearly, the "stop and forget" approach suits the psychological need of the inexperienced and naive trader to somehow "remove" a "losing" experience (with all the possible "pain" that can be associated to that). Instead, and this may be a bit counter intuitive, the rational manager must acknowledge and embrace this information, and use it for the future actions. And not only embrace it psychologically, but also make it an integral part of the automation, and therefore, even the simulations used to calibrate the games.

    ERY_7.png

    (In the meantime, considering the many buy players "locked" in loss due to order rejections, I have taken a good position of some ERY puts (400), in order to "zip" some of the short position, and increase the leverage, to possibly facilitate the recovery.)
     
    Last edited: Oct 31, 2014
    #96     Oct 31, 2014
  7. End of the week. Still some remarks on the stored information. Let's note that actually the G-L (gain - loss) curve (dotted-green in the picture) gives a useful "overview" of the past trading information stored in the players. This is clearly additional information that could not be derived by the equity (PNL) curve alone (in cyan in the picture) and it clearly provides great and useful insight to the manager, especially in difficult times. A manager looking only at the PNL curve alone (so not in possess of the player information) would normally have little or no clue at all about what is going on and, when there is a drawdown, basically "freak out" in panic, often on the sudden realization that the approach used was delusional and based on a mix of imaginary "signals" and curve fitting (and further often having daily dreams and nightmares on an impending "armageddon"), he would actually have absolutely no clue about what to do next, and, under pressure, in most cases take the loss and leave the money on the table (then a post-traumatic period will follow). And this is the standard mechanism through which eventually most traders lose money. We can, instead remain in full control, read exactly on our G-L curve what is going on in the scalping activity. And, if we need more detail, we can look at the open players to possibly adjust our games.

    PNL103.png

    Basically, in the "transition" phase where the broker is continuously rejecting our orders (to either close the open player which are still positive, or the open new reversion players), what the application does is to either place hedging orders (allowed by the broker) or to continuously reduce the order sizes (in exponential progression) of the scalping orders, until the orders are executed, and then starting again the order size reduction loop. This essentially means that we are actually using also the margin information the broker is "feeding us back" through the rejection, to adjust our action. This means that, at a certain point, it is the broker himself to somehow provide a "piece of the strategy", because if the prices continue to go unfavorable, only hedging orders will pass and we have a partial recovery due to position inversion; if instead the price goes favorable (start going along the long term drift) a small portion of the hedging players will be closed by the application through automatic reduction of the orders, and in any case, wherever the price is going, it will continue scalping something. If the price reversion continues, this effect increases and we get closer and closer to the situation where we can scalp again without any rejection. In practice, this means that there will be a period where you will be continuously receiving rejection messages, and that is completely normal, and perfectly dealt with by the application. (This somehow intuitively resembles a die hard fighter in an extreme street fight against a group of attackers puts his back against a wall and start blocking and countering hard from there.)

    Further, looking at the open players, we can have a clear idea (both in terms of price direction and position sizing) of what we need to recover the instruments. In general, in this case we have: players "locked" in loss (in this case, mostly due to the margin method issue), and they will need the price to get near to them to "unwind", while in the meantime we will use their "size" to automatically increase the size the opposite side players, to continuously counterbalance their "effects" (even if virtual).
     
    Last edited: Nov 1, 2014
    #97     Nov 1, 2014
  8. jcl366

    jcl366

    Why do you think that other peoples' trade methods are curve fitted delusions, but yours is not?

    This is not for mocking you, I mean the question serious. I've seen that you started new journals here several times - this is the third or fourth - and always commented your large losses with remarks like "all is fine and goes as planned". Should that not meanwhile tell you that not all is fine with your trade method? You're doing a sort of grid trading, but maybe not in the right way?
     
    Last edited: Nov 1, 2014
    #98     Nov 1, 2014
  9. Hi jcl366,

    That's an interesting question, and I appreciate the way and politeness you put it.

    Your phrasing is however a logic inversion of the actual process.

    You must understand that my journey in trading is older than my older daughter and therefore I myself have been growing up, along with the development of an application which is meant to grow with me and incorporate my experience.

    Growing up means trying many thing, maturing, throwing the old things and look at them with a smile. Feeling the enthusiasm for new findings, throwing them away, and so again an again in a neverending journey, until our dimensional adapter (body and brain) on this dimension still works :)

    So formulating the question implying that I am biased towards something because this something is "mine" is actually a misrepresentation of the actual process. In fact, it is exactly the opposite. Something is "mine" (in the sense I currently do it) because is what is left after having thrown away all what is obviously a fallacy.

    When I mock something, it's not meant to insult anybody. I am actually insulting a "past version of myself", if you like. My intent is actually to provide some hints for those who are at a relatively earlier stage of their journey. I am also perfectly aware that many of them will not to be "ready" to understand what I suggest (just as I would have not understood it several years ago ), however it's a seed which will help them as they go along with the journey.

    Finally, please not read in these words a condescending attitude: I am myself in a journey and always ready to throw away something which is conceptually fallacious. I just get a bit nasty when I sense scammers and people who deliberately try to deceive others or cause damage with their idiotic behavior.

    About the "all is fine and goes as planned" you are probably referring to the previous thread, where we had the "Zorro" idiotic troll ("rob the rich to give to himself", and "please give me a gift if you by chance should make 20K") intervention in the middle of the load phase along other poor idiots, and the withdrawal of the investor at just 20% drawdown.

    I invite you to go back to the thread, and you will see we had reached the peak of the "investment" with no significant loss stored in the players. Take a look at the GDXJ and NUGT charts from that peak and you will readily understand that with that kind of load we had, our investor would now be happily sitting on a millionaire account. Is a 375% move good enough for you (green lines are spaced 1%) ?

    NUGT
    NUGT_375.png

    Instead, we had a few idiotic trolls start screaming and denigrating at a mere 20% DD on a strategy based on drifting leveraged instrument with a structural monster decay. This I consider actually criminal, especially because in that case there was other people real money involved (and a huge work from my part). And these are immature losers with no respect for other people money and development efforts, which they cannot even start imagining.

    I say that if you "freak out" at a 20% DD you cannot trade at all, because you have lost money even before starting trading (please, note I am not saying this happened in our specific case: the investor might have his personal reasons to use liquidity elsewhere). It's like doing a backflip: if you have fear and hesitate just at the moment you turn back, it's likely you will break your neck: once you start you must perform the move with decision through the end, or you better not start at all. DD is part of any strategy: of course you must be in control and understand its reasons (and not be a DD caused by random entries = entries based on illusory signals). The problem is most people have DD whose nature they do not really understand, because their strategy is based on made up signals (therefore essentially "random entries" from a larger perspective), and that is why they panic, when things go differently from their ("curve-fitted") simulations on past data. The only reason why they adopted the strategy is because through twisting parameters and adaptation in endless backtests the computer was obviously able to generate a curve-fitted version of the strategy (that is what computers do best). But that is a mere interpolation exercise, like those I can ask my student to carry out as homework, but have no relevance for real trading. There is no reason in the word why the interpolation could be extrapolated. And in any case if you ask anyone, he will not be able to utter a single sensible and rational reason.

    The approach I am currently using is not based on "signals", or any attempt to adapt to past data but it merely relies on hedging capabilities and capital, maximizing their action. There is no past data involved, except the trading information I create myself, and of course structural data (drift, etc.), whose nature is objective and of "mathematical" nature [which of course if one is willing, might even be replaced by some predictive analysis].

    So the point is not "methods are curve fitted delusions, but yours is not?". But the exact opposite: they are not "mine" because they "are curve fitted delusions".

    Can you blow up ? Sure thing. But if you are willing to embrace some drawdown (I personally would be willing to go to 100% with my own "risk capital"), have decent capital and suitable account with suitable margin method, have some patience to wait good moves, protect with options, and do the right automation (right software implementation), you can be actually "larger" than mkt, and systematically extract the "deserved" profits ("deserved" because you had patience, capital, sustained some DD and had the right technology and method). The idea of making money from nothing based on "magic signals", is something I cannot accept, but the world is beautiful because it's diverse and we will always have Hitchens and Dawkins types along "creationists" and religious fanatics, and in this small space of my own thread I think I have some right to put forward my personal point of view, and I welcome any debate, that is not based on gratuitous insulting, and invitations to move to agricultural activities, but on a fruitful discussions of ideas and concept using rational arguments. Of course people easy to insults, just demonstrate that they have no rational arguments, and no education, and it's therefore likely they will never make a dime in the mkt.

    (Ah, btw, about "grid trading" has nothing to do with what we are doing here. The green "grid lines" I draw on the screen are just helpful to understand the relative amplitude of price moves. A utility for the manager.)
     
    Last edited: Nov 1, 2014
    #99     Nov 1, 2014
  10. jcl366

    jcl366

    If I understand you right, your current method is the result of a selection process during which you've thrown away many other methods that you now consider curve fitted delusions. But why have you thrown them away, but keep the current one, despite the losses?
     
    #100     Nov 2, 2014