Not replying to anybody in particular, just responding to some points that have been made. Some very thought-provoking posts in this thread.
It’s easy enough to pick ‘real’ charts from randomly generated ‘fakes’. Sure, the fakes exhibit the same kind of patterns (reversals from pinbars, head’n’shoulders, channels, occasional respecting of fibo levels, etc etc) merely fortuitously. But they don’t have the spikes caused by high impact news. Or weekend gaps.
Agree that ‘random’ is a nebulous term. I like PipMeUp’s 2-dice analogy — betting on a higher probability total will give you an advantage (edge) over betting on a lower probability total, but the outcome of each individual roll nonetheless remains random. Whereas a (fair) coin toss is 50/50, and no edge is possible. If markets are always random in the way that a coin toss is random (I assume this is what is meant by ‘random walk’), then every trading methodology has a mathematical expectancy of zero (ignoring costs), and all trading performance, even long term, is fortuitous. But if markets occasionally offer situations of directional or behavioral bias — similar to PipMeUp’s dice — then there is the potential to exploit these situations for profit. So while I can’t forecast each individual dice roll, I can nonetheless perform an implicit ‘forecast’ (e.g. that a 2-dice total of >=6 or will occur more often than not) on the basis of probability.
At their base level, markets are anything but random, because they represent the collective sum of every participant’s decision, and there is a rationale behind every one of these decisions. However, to whatever extent the sentiment and agendas of the heavyweight participants remain unknown (and/or the trader's inexperience at reading whatever useful info is provided by a price chart), price movement will appear to be random. Hence ‘randomness’ here is really a misnomer; it is merely reflecting the lack of information that the trader must work with.
I frequently read statements like “S/R doesn’t work”. But the fact remains that there are certain key levels where heavyweights have placed large volumes of orders, causing price to stall — and very predictably so — at these levels. In other words, clear evidence of non-randomness or behavioral bias. But I respectfully submit that the question is not so much whether this bias exists, but the extent to which it’s possible to devise a system that’s capable of exploiting it, beyond costs, on a probability basis. Alas, my br0ker won't allow me to profit from a bet that “price will stall here for a while” or “this news announcement will cause a spike”, even if such a prediction was to be correct (say) 98% of the time: a massive edge. Hence the point I’m trying to make is that, while aspects of price movement may frequently be extremely predictable, the fact that one must devise a methodology involving entries and exits that catches a move of sufficient length, on average — to overcome not only costs, but also untimely stop-outs, and other hazards imposed by a ‘conventional’ methodology — is an entirely different (and much more difficult) assignment. The trader is required to get both the ‘direction’, and the ‘timing’ of both the entries and the exits, accurate enough, on balance, in order to succeed.
Profits are reduced, and losses increased, by the inefficiencies that are inherent in the conventional trading paradigm. For example.....
Entry: the trader can either attempt to anticipate a bounce (‘catch a falling knife’), or wait until there is some evidence of the bounce. If the former, he runs the risk that he will be wrong much more often than he is right, reducing his win rate. If the latter, he is forfeiting pips, with the result that he catch only a lesser part of the total eventual move.
Profit Exits: traders tend to use either TPs and/or trailing SLs. TPs leave pips on the table, and trailing SLs forfeit the number of pips by which price is trailed. Either way, the trader again catches an even smaller part of the total eventual move.
Loss Exits: regardless of whether the trader uses a hard or mental SL (or a ‘hedge’ to neutralize his position), many would-have-been-winners are not only going to be curtailed prematurely, but losses will result instead — a double whammy (how many times do we find ourselves cursing a situation where we got stopped out by 1 pip, from a move that subsequently would have made 100 pips?). But if the trader doesn’t employ some kind of loss exit, he leaves himself open to the possibility of unlimited loss.
Costs: as if all of the above wasn’t enough, every order incurs costs, reducing the size of every winner, and increasing the size of every loss.
In summary, markets can trend, and occasionally very ‘predictably’ so, but (1) the trader might only be able to workably catch a very small part of that move, and (2) find that many ‘winning’ trades unfortunately were first stopped out as losses. In other words, the trader’s methodology unwittingly works against him as much as any ‘randomness’ in the market does. Difficult though it may be, finding non-randomness in price behavior might even be the easier part of the battle.
It’s easy enough to pick ‘real’ charts from randomly generated ‘fakes’. Sure, the fakes exhibit the same kind of patterns (reversals from pinbars, head’n’shoulders, channels, occasional respecting of fibo levels, etc etc) merely fortuitously. But they don’t have the spikes caused by high impact news. Or weekend gaps.
Agree that ‘random’ is a nebulous term. I like PipMeUp’s 2-dice analogy — betting on a higher probability total will give you an advantage (edge) over betting on a lower probability total, but the outcome of each individual roll nonetheless remains random. Whereas a (fair) coin toss is 50/50, and no edge is possible. If markets are always random in the way that a coin toss is random (I assume this is what is meant by ‘random walk’), then every trading methodology has a mathematical expectancy of zero (ignoring costs), and all trading performance, even long term, is fortuitous. But if markets occasionally offer situations of directional or behavioral bias — similar to PipMeUp’s dice — then there is the potential to exploit these situations for profit. So while I can’t forecast each individual dice roll, I can nonetheless perform an implicit ‘forecast’ (e.g. that a 2-dice total of >=6 or will occur more often than not) on the basis of probability.
At their base level, markets are anything but random, because they represent the collective sum of every participant’s decision, and there is a rationale behind every one of these decisions. However, to whatever extent the sentiment and agendas of the heavyweight participants remain unknown (and/or the trader's inexperience at reading whatever useful info is provided by a price chart), price movement will appear to be random. Hence ‘randomness’ here is really a misnomer; it is merely reflecting the lack of information that the trader must work with.
I frequently read statements like “S/R doesn’t work”. But the fact remains that there are certain key levels where heavyweights have placed large volumes of orders, causing price to stall — and very predictably so — at these levels. In other words, clear evidence of non-randomness or behavioral bias. But I respectfully submit that the question is not so much whether this bias exists, but the extent to which it’s possible to devise a system that’s capable of exploiting it, beyond costs, on a probability basis. Alas, my br0ker won't allow me to profit from a bet that “price will stall here for a while” or “this news announcement will cause a spike”, even if such a prediction was to be correct (say) 98% of the time: a massive edge. Hence the point I’m trying to make is that, while aspects of price movement may frequently be extremely predictable, the fact that one must devise a methodology involving entries and exits that catches a move of sufficient length, on average — to overcome not only costs, but also untimely stop-outs, and other hazards imposed by a ‘conventional’ methodology — is an entirely different (and much more difficult) assignment. The trader is required to get both the ‘direction’, and the ‘timing’ of both the entries and the exits, accurate enough, on balance, in order to succeed.
Profits are reduced, and losses increased, by the inefficiencies that are inherent in the conventional trading paradigm. For example.....
Entry: the trader can either attempt to anticipate a bounce (‘catch a falling knife’), or wait until there is some evidence of the bounce. If the former, he runs the risk that he will be wrong much more often than he is right, reducing his win rate. If the latter, he is forfeiting pips, with the result that he catch only a lesser part of the total eventual move.
Profit Exits: traders tend to use either TPs and/or trailing SLs. TPs leave pips on the table, and trailing SLs forfeit the number of pips by which price is trailed. Either way, the trader again catches an even smaller part of the total eventual move.
Loss Exits: regardless of whether the trader uses a hard or mental SL (or a ‘hedge’ to neutralize his position), many would-have-been-winners are not only going to be curtailed prematurely, but losses will result instead — a double whammy (how many times do we find ourselves cursing a situation where we got stopped out by 1 pip, from a move that subsequently would have made 100 pips?). But if the trader doesn’t employ some kind of loss exit, he leaves himself open to the possibility of unlimited loss.
Costs: as if all of the above wasn’t enough, every order incurs costs, reducing the size of every winner, and increasing the size of every loss.
In summary, markets can trend, and occasionally very ‘predictably’ so, but (1) the trader might only be able to workably catch a very small part of that move, and (2) find that many ‘winning’ trades unfortunately were first stopped out as losses. In other words, the trader’s methodology unwittingly works against him as much as any ‘randomness’ in the market does. Difficult though it may be, finding non-randomness in price behavior might even be the easier part of the battle.