I was all this week in the market, in the side of the trend and ended the week 1,22% up. Remember, I am always with the trend and with very small size. I have been sticking with the trend for a long time. My best trades are exactly these. Most of what I lost was for trying to predict the market, entering big and being scared out. Thanks for the link anyway. I will try to check it out. Is there a pdf? There are many possible trading strategies. I think no one detail is a complete shit in itself. You can build a profitable trade plan that begins with tossing a coin. The plan as a whole is what matters.
I don't know if there's a pdf but this book is worth its weight in gold. This book completely changed my perception of the market and trading in general. Truly a great book!
Martingale is by definition a losing trade. What you should do is kelly bet. THe larger your edge, the more money of yours you should risk. The math is very simple...finding how large your edge is or even finding an edge is the hard part. Going heavy when the market will move your way is bullshit...it only moved your way after the fact.
Hmmm...not like you to hold back like that. Tell us how you really feel about negative progression sizing!
$1,442/$3,300 = 0.44 At his one micro lot size in the EURUSD his apparent Kelly fraction of 44% implies a not completely unreasonable risk-adjusted edge over the risk free rate for his always-in trend following forex system. At a fraction of 0.44, Kelly is a negative progression system. When you lose, you increase size, when you win, you decrease size -- for a uni-directional always-in system Kelly fractions under 1.0 are a martingale. Wiki cut-and-pastes notwithstanding, in common usage the term "martingale" means a negative progression betting system. You are simply suggesting that the OP subsititute one form of optimal martingale (Kelly) for his own suboptimal martingale (classic Small Martingale). Also, non-kelly negative progression martingale-type systems are proveably optimal under certain circumstances, particlarly when you can convince someone to pay you to play (think Nick Leeson), and are in common usage.
On the contrary, if your compensation is determined by common risk-adjusted return measures -- e.g. Sharpe, Sortino, Jensen's Alpha, etc -- on a discrete sampling interval (monthly, quarterly, yearly...), then sub-interval martingales make perfect sense. This can be proved analytically or through MC simulation. Use this simple rule: If the Sharpe is above your benchmark in the first half of an interval, then decrease exposure for the second half, if below in the first half, increase exposure. Test this, you will see martingale magic at work.
Calculating and trying to increase probabilities is exactly why I started this thread. I simulated and used in a small live account for a short experimental period a mantingale EA that doubled each 20 pips and stopped out at 200. This is suicidal. It accumulated nicely for 1 to 2 months. Then with 2 big losses it gave back all it had gained. It is typical, I know. But at the same time it was having the first loss I ran a parallel mantingale in another account (obvious, very small lots). While the EA lost in one account I won in the other. It was a strong pull back within a trend. That's the trick. Never go against the trend. When the trend resumed I was quickly back to good profit. If the market reverses, there will be losses, but that is the role of trend monitoring. If you are not too big at that point these losses will not take back all your profits. I think it is a good idea to decrease progressively in size as the trend grows older increasing the odds of a reversal. I would like to know if someone has some statistics about this. I know a grid of 20 pips is suicidal. 305 pips is not necessary. Maybe something around 50 to 150 pips catches good pull backs without much pain. Only for reference: 300 pips on 0,01 lots represents less than 1% drawdown on equity. Extremely, extremely, extremely low risk.