Page 1 of 1

Trading myths and statistics

Posted: Wed Dec 17, 2008 8:03 pm
by Htarlov
I have an idea - based on Discovery program "MythBusters" - to make a topic (or some area) on forum in which we would take up on popular trading ideas and would disprove (or prove) them - with use of math and statistic.

What do you think about that?

Or is there something like that present on this forum in some place?

Posted: Wed Dec 17, 2008 10:45 pm
by eudamonia
Htarlov,

Interesting concept. Don't think anyone has done that on the site. The problem with "myth busting" and trading is that most of the trading truisms are true most of the time - until they're not.

Take the common truism that trading "with the trend" (whatever that is) is a good idea. But there is nothing quantitative about that statement. How do you define a trend? What market(s)? What timeframe(s)? Should you trade breakouts or pullbacks? How does money management and scaling in/out play in? Is there a minimum return/drawdown ratio that could define that trends do/don't work? What about number of trading opportunities? Is there a limit to the scalability of trading in such fashion? What about psychology?

In other words how can you "prove" or "disprove" something so nebulous in the first place? More importantly is there any value in doing so?

Edward

Posted: Thu Dec 18, 2008 4:22 am
by TheRumpledOne
Trading is about making money.

That's what the focus needs to be on.

Posted: Tue Dec 23, 2008 4:22 pm
by Htarlov
TheRumpledOne wrote:Trading is about making money.

That's what the focus needs to be on.


Yeah, probably you are right.

Busting trading myths and giving that information for the public is probably counter-productive thing for a good trader.

The more noise-traders or nonsense-strategy-traders the better the market is for those with good working strategies and money menagement. :twisted:

And disproving some concepts is probably pretty hard (because they are not very good defined nor used the same way by many traders).

But for a self-use testing strategies and "myths" statistically isn't so counter-productive. It's better to have a strategy that you know that works, will work (at least for some time - because when many people will get the same ideas and use similar systems they will change the dynamic model of the market and those strategies will stop to work so well), is stable and you know why it does work.

I made a post about myths because in many places I see information about for example Fibonacci trading and Eliot waves and this seems like some strange hocus-pocus method to me (but I'm just a programmer interested in trading, science and few other things and I work in pretty young company that specializes in statistic arbitrage). It looks like horoscope-trading for me.

I'm interested if someone did a statistical search in that area (and I'm thinking of doing some search by myself).

I know there are traders that sucessfully use Eliot waves.

But with proper money management pretty many strategies should work for a pretty long time for some percent of traders - even making some of them rich - just on luck basis (it's like using well known strategy for roulette - you can "stretch" risk in time so the probability of loosing much is low in a short period and probability of winning is high, but on a longer period there is a high probability of big "blow up" - you can get big "blow up" rarely with strategy of money management for roulette or small looses often without it).

Most of traders use more than one indicator and technical of analysis (including subconscious and conscious patterns recognition made by their brains) - so maybe Eliot waves and fibonacci sequence isn't statistically significant (it's pretty probable that price will change direction near one of the levels of retracement but it often doesn't, it's for me pretty hard to believe that it's statistically significant - but maybe on some degree it is because many traders use it and it can be self-fulfilling-wish), but maybe it's needed and good for other purpose?

For example for timing - to choose a place where the decision should be made?
Or for making question about something continous (where will be the price direction change) into question about something quantitative (what level of fibonacci retracement)?