cesko, yes, good point. randomizing the initial short/long decision will cancel out the market trend real nicely, so the trades will be perfectly 50%-50% symmetrical (if we ignore the commissions that is). it's really easy to prove that. - jaan
DH, I know you feel that way, but, FWIW, jaan/cesko are making arguments as they would be made by Mathematicians (these arguments are elementary, but probability is notorious for non-intuituve results.) They are exploring boundary conditions, etc., and working their way from there. I agree with you that they won't get anywhere, but it does not render the mathematics "pointless." BTW, I thought all swingtraders were morons (ay ay yai, the heat I will get) until I read the way you approach it. Very elegant, I will think on it further... nitro
Darkhorse It still won't work. I wonder why? It's easy to just say that. Not that it matters one whit anyway. You are absolutely right. What a pointless argument... Yes but there is not much to talk about anyway. I mean tradingwise. Let's see: Uptrend: higher highs, higher lows Downtrend: lower highs, lower lows Sideways: support, resistance Liquidity Volatility All that changes are the tools.
yes it would -- it's a simple mathematical fact. effectively, after applying the random short/long decision, each trade that would have been a winner has a 50% probability of ending up a loser, and each trade that would have been a loser has a 50% probability of ending up as a winner. hence, all trades will have a 50%/50% probability of ending up as winner/loser. QED. hehe, i can tell you're a discretionary trader, right? for us, system traders (i trade a 100% mechanical system with no discretion whatsoever), probabilities and expectancy are the only things that matter. - jaan
Jaan, This still does not mean a 50% chance because if the market is weak and few stocks are making a 10% move to the upside, you don't have a 50% chance of getting out the way you entered. You are acting as though stocks have an equal chance of going 10% either way. Say you make 100 trades. Assume 50% of your trades are sells (since the entry is random) and the other 50% are buys, and the market is weak. Say you win on 70% of your sells (35) and only 6% of your buys (3), you would only have 38 wins, not including commission. And since you don't ride winners, but rather give 10% each way, you are losing a lot of money. This assumes all the stocks are the same price. Now consider that 10% of the value of stocks that are at different prices, and you have one more variable that muddles this system. By the way, what system trading do you use, a black box?
so what? if you toss a balanced coin 10 times, and happen to end up with 10 tails in a row, would you say that the probability of a tail for the balanced coin is 100%? the point is (and i believe that's also close to magna's original point), having a 50% chance of a win does not necessarily translate into 50 wins in every 100 trades. it also works the other way around: if you have 38 wins out of 100, that does not mean that your winning probability is necessarily 38%, as you seem to imply. in your example, you still had a 50% probability of win (see my proof in the previous post), yet you were just unlucky to only end up with 38 winners in 100 trades, because only 6% of longs ended up as winners, not 100-70=30% that would have been more probable (ie had the market bias affected longs and shorts similarly). i hope this made sense -- as previous posters have said, the probabilities can be surprisingly tricky (and surprisingly powerful, too). no, i'm a partner in a software house, and we trade a proprietary system that we have developed ourselves. - jaan
Nitro: In some ways that is part of the problem, and it is a problem that crops up again and again in economics. You can't apply textbook math to a chaotic environment, there are so many hidden variables and textures that the number crunching is usually drowned out by any number of a hundred hidden factors. This is precisely why economics as it is currently taught in universities is 99% crap/waste of time- they are trying to apply cut and dried robotic formulas to human decisions, which does not work. Also I'm perfectly happy for folks to think of swingers as morons...the less peeps in my timeframe the less competition I have to fight with....
i can see the discussion taking the "discretionary vs. system" path again, that has been done numerous times here on ET, so let me give you a summary of the argument that will follow: joe discretionary: "you can't apply a mechanical system to market, because there are too many hidden variables and unforseen events that the system is unable to take into account." bob system: "i don't need a system that is able to see every detail there is, to predict every event there will occur. (if i had that, i would make all the money in the world, in, what, a week?). instead, the job of a system is to give me a small edge in the long run -- you don't need an incredibly sophisticated system to do that." - jaan
? I never disparaged systems, in fact if you look back a ways you will see that I have defended systems (in this very thread I think, or if not this one then another very recent one). My point had nothing to do with methodology, but rather with the futility of trying to precisely target percentage results as applied to a chaotic environment with dozens or even hundreds of factors that aren't factored in to the "equation" (and chaotic does NOT mean random FYI). You can get an idea of where the chips will fall but you cannot say "I am assured it is possible to get 50% win/loss" or XXX reward to risk or any such thing. It is the false precision I disparage, not the validity of a logically grounded mechanical method. To borrow from Bobby DeNiro, when you "assume" your calculator always tells the whole story, you make an ass out of You. Period.