What BS? It's the indicator developed by Proflogic. You can read about it in his thread, where he tells what prime number charts he uses and there are multiple indicators that you can download from that thread to try for yourself. I am only refer to this indicator to explain why I once used prime number charts as an answer to a question. Next you will ask me why is the sky blue? How about you try Google instead.
Let's summarize a little. 1) Why can't he use a volume chart? Because like a 1 min chart, you get more noise. I then went on to explain that a higher prime number chart would be better to use if you don't want to use a time based chart. 2) Why is a prime number chart better to use than a non-prime number chart. Because it worked with an indicator and rules for a PA or TA system that I experimented with. If you are just going to randomly do shit, and then you can't figure out why you can't spot trades, maybe it's because you are not using a system and following the rules for that system.
I am pretty sure no-one here (or anywhere) can prove through a 1000+ trades backtest that a prime-number (P) based volume chart is statistically better than its two neighbors, P-1 and P+1. Or even the closest 2 multiples of 10. A lot of opinions so far, but nothing backed by any statistical analysis. It isn't that difficult to find price-action based patterns with an "edge", but there is no way to escape the statistical analysis (anything less than 1000 samples is ultimately a waste of money, IMO). And no, there is no way to outsmart the market and have a 100% probability on the outcome of any trade. So instead of wondering why the market in this *particular* instance did this or that, try to find-out what the market does more often than not over the course of 1000+ similar situations. Also, forget about making the next trade a winner - just concentrate on making the total P&L of the next 100+ trades positive.
No offense intended, but that doesn't prove/show that a prime number based volume chart is "better" than any other vol chart. Having spent literately thousands upon thousands of hours looking at and studying charts of all types markets and time frames, I have to agree.
Backtesting is useless due to similarity of the orthogonality of the probationer,hence,this line: "Also, forget about making the next trade a winner - just concentrate on making the total P&L of the next 100+ trades positive.'
In my experience the only people who say "backtesting is useless" are scammy system vendors and "gurus" who don't want to acknowledge that their systems backtest unprofitably because then their followers would move on to someone else. This goes double if the scammy guru is charging money for "instruction."
Back testing is very important for a simple fact that if the system fails it, no way it will be profitable live. That being said, many systems that pass back test, fail in real live due to execution, slippage, commissions just to name the few.
I'll agree with that. You need to factor in commissions and slippage. And make sure it's not curve fitted. And I'm sure there are some systems that would be very hard to backtest, such as very short timeframe systems, ones that use data that is difficult to collect historically such as DOM data, etc. But still, the most adamant "backtesting is useless!" people are almost always the vendors (who have systems that don't backtest profitably).