Everything is completely dependent on the next datapoint with this crowd. A couple of weak numbers and/or the smallest selloff in shares and QE is right back on. Borderline sociopaths.
I wrote about Groupon here more than 1 year ago, called them a scam who were engaging in lending transactions with junk companies. This guy seem to agree http://venturebeat.com/2012/03/31/why-groupon-is-poised-for-collapse/#.T3jAd6hKVbo.twitter Problem is, it costs 30% a year to short GRPN. That is priced in on the SSFs(They sell at a discount) and I believe it affects the options as well So its kinda expensive and it doesn't leave much room for error, but it might be the correct thing to just go ahead and short it. Stocks sells at 5x revenue
I disagree, let me explain. Firstly, current risk/reward is judged purely by your current exposure and the current trade expectation - earlier trades and P&L have nothing to do with it. E.g. if you are long 30% of capital in market X and just entered, then you have identical risk to someone who made a superior entry earlier on and earned some profits as a result, but is now also long 30% of capital in market X. As Paul Tudor Jones said, there is no such thing as 'the market's money', and your risk today is from yesterday's closing prices, not from where you entered the trades. So, having on an earlier position, even if profitable, doesn't reduce your risk once you put on the big size. Secondly, unless you are in an opaque or illiquid market where you can't tell the price without doing actual trades, having on a position gives you zero additional information compared to simply doing the same observation with no position on. You can see the same price action, ticks, fundamentals, valuation data, news etc with a flat position as with an outright position. No possible data that you could observe whilst having a position on, is not also observable by someone else who is flat. In fact, for many (inferior) traders, being flat lets them think clearer and avoids the common psychological biases that come from having on a position. Thus there is no such thing as a 'test' position in a liquid transparent market unless your trading is so large that it moves the price (e.g. the old plungers who would sell heavily to see if demand was solid). There are simply phases of a trade where the trade expectation has different odds. The poker analogy is flawed. First, in most cases you must play and pay on earlier streets in order to earn the right to see the later streets where the big profit opportunities usually arise - in fact, if you don't bet enough, the pot may be too small when your payoff arrives. Early bets are therefore 'investments' as you point out. In trading it is totally different - betting early on makes *zero* difference to your ability to bet later on, in fact it may compromise it (e.g. if your early position loses money - such as when buying a market crash before the bottom occurs). Secondly, in poker you are playing a small number of opponents, and their responses tell you something about what cards they may hold. They can be manipulated by your own betting, or make psychologically-induced mistakes like getting married to a hand. Your early bets thus give you information about about your likely odds, information you could not have got without betting. In the markets, you can get that information whether you bet or not. Therefore there is no informational advantage on the typical trade to taking an early position. The poker players early bets are both for +EV on the bet, and for information advantage - the latter often being more important. The trader's early positions are purely for +EV, there is no informational advantage gained from them. This is the problem with trying to argue by analogy - there is no actual justification to it, and it tempts people into drawing unsupported conclusions. It is far better to just use pure logic and concentrate on the specifics of the situation at hand. About 90/10 ratio - it occurs because occasionally a trade comes along that is far superior to the norm, and good traders then bet big on it. The other 90% are worthwhile only to the extent that they are profitable, or provide information advantage. But as I pointed out, it is rare that any informational advantage is gained by having on a position, that cannot be gained simply by following the market as closely as if one had on a position. So, the 90% of trades are generally worth it purely for their moderate profitability. My hypothesis is that they may actually be net losers once you take into account the distraction of focus that comes from placing and managing marginal trades, rather than devoting 100% of your time and energy to finding and optimising the 10% of home run trades. I can't prove this last point, but anecdotal evidence suggests it is worth investigation: if you read what these traders say, none of them ever say they traded too infrequently - they all say they overtraded if anything. Preservation of capital + betting big when a home run sets up is the way to superior returns, at least on macro and other 'fat pitch' +gamma strategies. The only way this would be wrong is if there was no way to tell in advance the 10% from the 90%. But if there is no way to tell, then you must logically bet the same size for all trades. Yet none of the great 'home run' style traders ever did this - they all vary size massively. It is utterly irrational to vary size massively if you think all trades have the same observable odds before you put on the position. Therefore, the 90/10 traders themselves implicitly state that odds vary significantly and that they can observe this before they place a trade. What pays the greater dividends - a marginal hour spent on a marginal trade? Or that marginal hour spent on a trade that may make your year or decade?
That's a good point, you have to adjust with the payoff ratio, not just the odds of success. However, I'm not sure how you would be able to justify it in advance by any kind of theory of markets or trading, and back-testing with such a sensitive output would be very prone to curve-fitting. This kind of strategy can work for a while and then stop working, and there is no real way to tell if it's just a legit drawdown in the method, or whether the method was invalid/random all along. I think the history of markets and trading strategies has shown that robustness is much more sustainable and profitable over the long-term, than optimisation. I want to *know* (as much as possible) my strategies have sound reasons for working in future - not just that they worked in the past. It's like observing a dice that rolls a 6 more than average e.g. 30 times in 100 throws. Was it just a streak, or is the dice biased? Unless I can somehow measure the weight distribution and flatness/shape of the dice, I am not going to bet any significant sums on it. I want both good past results, and good theoretical reasons to expect those results to continue. That's why I'll never bet on the Super Bowl theory, dogs of the dow, january effect, or other correlations with weak theoretical justification.
Although this is a decent method for beginners and intermediate traders, I don't think it's rational to trade or not solely based on what your past trading results were. For example, let's say I am long a bunch of stocks, then Iran gets invaded and the market falls 10%. I then see an amazing opportunity to put on a once in a decade trade in the oil market, with minimal risk, huge chance of success, and a 10:1 reward to risk ratio. Should I pass up this gift just because I lost more than my monthly stop loss due to a hard-to-foresee event? No. I should not hamstring myself by artificial constraints. I should do the ONLY rational thing to do for any trader - try to maximise my current trade expectation per unit risk, within my risk tolerance boundaries. That 10% has already been lost, I now have 90 units left - my task is to deploy them as optimally as possible. And that means taking the oil trade on reasonable size. Remember, if you stop yourself out, you will either have to retire from trading, or start trading again at some point. There is no reason to expect that your P&L will be any better if you start in 1 month than if you start again today or tomorrow. If you are mentally destabilised or otherwise incapacitated, then fair enough. But if you are of sound mind and are prepared, then you should trade. If you made any mistakes, analyse and fix them, avoid repeating them, then get on with your trading.
i do not agree. If you know that at current price market has 80% chance of reversal on a daily basis - swing trade, then playing long on a minute level or so only with 20 x smaller stop & risk - same size - will yield better results than loading the boat with appropriate stop. Of course your target is 50 x risk and hopefully you loaad couple of batches and of course have couple of failures as well.
There's one time it can work and has good theoretical reasons for doing so - when a stop got triggered, people sold on size, and then the market rebounds rapidly. This indicates that people got stopped out for BS reasons, and yet the market trend is solid. You then have a load of sold-out bulls who will have to buy back at higher prices (which is where you should book your profits). The fear of it 'being in the price' is not applicable here because you are entering minutes, hours, or a day or two after - people have not yet had time to react and reset their stops, and in any case when this pattern works the market has usually moved considerably away from the stop anyway. And if the low of the post-stop selling wave gets penetrated, then clearly the thesis was proven wrong and you will be correct to exit. This is a classic case of naive technical analysis based on nothing, versus proper technical analysis based on sound theoretical reasoning about market structure and the behaviour of speculators.
The problem being that we don't know what 'makes money' (i.e. will make money in future), we only know what made money in the past. Given that some past performance is clearly random/lucky/spurious, we need more than just past performance to justify employing a strategy in future. We need a reason to think that the strategy is based on more than just data-mining from the past. For example, the equity risk premium is not just observed in data across many markets in many time periods, it is also justified theoretically in that equities have higher volatility and inherent risks than bonds (generally), thus investors will demand greater returns to assume those risks. You have sound extensive data, and sound theory to support your expectation of future returns. In something like one optimised stop versus another, in most cases there is no theoretical justification for it - so, the chances are much higher that it will just be data-mining. Furthermore, the more sensitive something is to optimising, the more likely it is to be pure chance. A sound system of stop placement should work fairly well whether its 2xATR, 3xATR, 4 or 5x. Yes there will be some difference in results but if one works amazingly and the others suck, chances are its a bogus data-mined conclusion. The 'theory' you are using in stop placement is "intact trends shouldn't retrace too much more than X - if they do, the odds are high that the trend is broken". Needless to say, this theory doesn't exactly scream out for precise and optimised stop placement - it's much more vague and general. So, your stops should fit the theory - maybe have 3 or 4 stops spread out over a decent range (e.g. 2-4xATR), rather than trying to curve-fit to within a few ticks.
Well, by definition that is false - if the 20x smaller stop is more appropriate than the original one, then the latter is not the appropriate stop. The point is how to identify the appropriate stop, and my argument is that you need some sound theoretical reason for thinking that the stop placement chosen provides you with any kind of edge.
We really see things differently here. And funny you mention PTJ, because he is exhibit A as to the reason why... In Market Wizards, Jones laid out some basics of his trading approach. One key aspect was 'probing' a market for entry points, trading in line with conviction / fundamental bias but only risking very small amounts of capital until a position gained traction. The point of this is very simple: * When you win, you have the ability to win big. * When you lose, you tend to lose small. As far as exposure levels on a trade, "how you got there" is important due to the consequences of being wrong. A trader who starts a trade at 1% planned risk and then builds up to 30% under favorable conditions can be wrong half a dozen times in a row without great damage to his account. A trader who starts a trade at 10% planned risk (committing a large chunk of capital off the bat) would be taken out on a stretcher after a string of losses. The point is not to be comfortable with strings of losses, but to have defensive preparation as such that outliers do not hurt you. Keeping initial exposure small, then building that exposure quickly under favorable conditions is a very important aspect of building desirable asymmetry into the win / loss record (average loss quite small, average win much larger). Once again, we disagree completely here. This may be a matter of experience. In Trading Risk Ken Grant (risk manager to many of the top funds in the world) talks about how top traders frequently have "test positions" in the sense of needing to be "involved" to have a real sense of a market. There is a psychological aspect and an intuitive aspect to making "probe bets" where you have a small level of conviction in regards to a market, and want to develop a feel for how it is acting before dialing your conviction higher. This is also my natural and routine experience. Not true. Do you play poker for meaningful stakes on a regular basis? Because I do. And I have thought deeply about the theory, as intertwined with practice, because I enjoy it so much (and it acts as a mental sharpener). You do not have to "play and pay" on earlier streets to "earn the right" to see the later streets as you say. The analogy as I expressed it is exactly correct. There are certain times when it is possible to enter a pot very inexpensively prior to the flop, relative to stack percentage and implied odds. And once the flop comes, everything changes. This is precisely why it is theoretically correct to play hands like 3-5 suited in position under certain conditions in deep stack cash games, where it would not be correct to play such hands in a middle or short stack game. Read "Harrington on Cash Games" if you would like more on this. A discretionary trader who uses both technicals and fundamentals DOES gain information, and valuable intuitive confirmation and conviction development, through the establishment of initial small positions with meaningful implied odds. Again, if you don't want to take my word for it, read Ken Grant.