The CL/NQ experiment

[AAAintheBeltway]

The whole idea behind using a computer to do the back-testing is to remove the human aspect from the test.

Which one may want to do if he is backtesting an indicator-based system. But judging from your charts, this is not what you're trying to do (though you do plot an EMA which is preventing you from taken the proper entries). If you were trying to backtest such a system via computer, I'd say go ahead, even though you'd have to do it all over again at some point in the future, most likely some near point.

A system based on price movement does not require constant maintenance because price is self-correcting. But this has all been hashed out in the If You Could Go Back In Time thread, with NoDoji and me and a few others against the world.

If you want to "remove the human aspect" from backtesting, and by that I assume you mean that you don't want to have to interpret trader behavior in order to determine probabilities and make judgements regarding trading decisions, then you've chosen the wrong approach. The "human aspect" is what trading price is all about. If you have no interest in that, then plot a MACD and an RSI and a slosto and begin again.

If one sees that over time the maximum consecutive losses is X then a proper position size can be arrived at to avoid running out of money.

Not if that maximum happens to exceed itself, which it may well do since markets are unpredictable. Even so, with what does one begin again after his resources have been depleted by that maximum number of consecutive losses?

Remember that this whole notion of "positive expectancy" was developed by someone who couldn't trade profitably, but failing traders have latched onto it like ticks because it provides the allure of being able to pursue a losing strategy and yet eventually show a profit, sort of like the promise of the afterlife. Unfortunately for those traders, it's nonsense. If one doesn't have a winrate considerably better than that 30% so regularly touted, he's going to go broke. Even 50% will be a trial without spectacular trade and risk management, each of which is rare among traders and practically non-existent in combination.

Of course a Black Swan event can happen and the 2007 economic crisis is one example.

An example of what? That decline was easily predictable. In fact, I sold my house in Phoenix in spring of '08, and not due to indicators but rather plain ol' price charts.

To quote you, "Gold."

Gabe is half right. You can make money with a system that has a 30% win rate. In fact, that is the typical win rate of trend following systems, eg volatitlity breakout, moving average systems, etc. What he is missing however is that these types of systems rely on a handful of enormous winners. You are not going to get those winners if you are exiting every time price ticks against you, or at the end of the day for that matter. I studied intraday stock index systems for a long time and concluded that there were no commercial systems that were reliably profitable. None.

The other thing about low win rate systems is that they are best traded by someone who is trading them across a portfolio of commodities, not just one market. Like Cramer says, there is always a bull market somewhere. You might wait years for one to come to the one market you are trading. Even across a portfolio of tradeables, these systems have huge drawdowns, meaning you need a big bankroll and a lot of patience and trust.

If you are trading intraday, you need a decent win percentage. You also need a way to let profits run without giving back too much.

 

[Gabe2004]

The economic crisis was NOT a Black Swan event, not even for long term traders/investors. There was plenty of time to lighten up, hedge, or move into cash.

I should have been clearer.
The magnitude of the crisis was the Black Swan event not the crisis by itself.

> Scroll my chart(s) to the hard right edge at the time my trading day starts.

> Reveal one bar at a time until a valid setup appears.

> If a trade triggers per my rules, log the entry price and time.

> Manage the trade according to my core management rule (I start with one rule and include additional info on the spreadsheet as noted in the next step).

> Note the result (profit target or stop loss) and include additional columns for result if stop moved to b/e after N ticks favorable, MFE before price returns to entry price, MFE before price hits stop loss (unless an opposite position is signaled based on a new valid setup), and flexible S/R-based target.

I have attached charts of how I do my back-testing.
The 5min chart is the main chart and the 1min is used to find MAE and MFE.

Pa(b)st

I don't understand.

Gabe

 

[Gabe2004]

See the trading plan I've detailed in my threads and plot an EMA on the charts to see the differences in entries and exits

Of course there will be differences but that does not mean the using an MA is causing me to enter into bad trades.

If you can't tell whether price is going up, down, or sideways without a moving average,

I guess that was a rhetorical question.

If a moving average is helping you to succeed, there's no reason for you to stop using it. OTOH, if you're continuing to fail even with its use, then . . .

Let's say that I don't think that using an MA as part of my system is the reason I have failed so far.

Gabe

 

[Gabe2004]

If you are trading intraday, you need a decent win percentage. You also need a way to let profits run without giving back too much.

Personally I find that a higher W/L is the way to go (60-70% is good enough)

Gabe

 

[foozler]

Personally I find that a higher W/L is the way to go (60-70% is good enough)

Gabe

If I may, Gabe, ask you a question: How did you determine this figure? On what basis have you found 60-70% to be "good enough"? And is there not a big difference between 60% and 70%? Over the course of 100 trades, a 60% win rate would have 14% fewer winning trades than a 70% win rate (assuming the back of this napkin is working properly).

I do think you should read over nodoji's "higher power" post above, and really think about how you are approaching this section:

You say the choice is clear and you're working on it. You have been working on it for a long, long time and guess what? It's not working.

You have seasoned traders here who are willing to act as your Higher Power and who can restore you to sanity. (What you're doing over and over again despite the fact that it isn't working is insanity.)

Turn your trading and your will (ego) over to one of these powers greater than yourself.

Read The Disciplined Trader and use it as a framework for making a fearless, honest inventory of your bad trading habits and your defects of character in general.

Admit to yourself and your Higher Power (just post it all here on this journal) the exact nature of your bad habits and defects of character.

Then move forward with the hardest part of all:

Become humble and willing to do the work suggested to you by the Higher Powers here even if you believe it's boring, unnecessary, of little value, doesn't apply to you because somehow you're [different, better, smarter, unique, special].

If you do all this hard work and end up finding yourself in the coveted position of trading a positive expectancy plan appropriately for several days, weeks, or months, then you have to be forever on guard against the first lapse in discipline. Your equity curve early in the journal is a graphic example of what happens when you don't heed Pa(b)st's warning long ago here on ET:

 

[Gabe2004]

How did you determine this figure? On what basis have you found 60-70% to be "good enough"?

It was an off hand remark to the following

I've also posted a detailed trading plan with a winrate of 80%

If I remove the 4 days in which I totally lost control, my win rate would be just shy of 62%.
This number, give or take 1%, has been my win rate for years.
The large losing days were/are what prevented my equity to grow.

I don't think that there is any win rate that is a correct number.

Gabe

 

[dbphoenix]

It was an off hand remark to the following

If I remove the 4 days in which I totally lost control, my win rate would be just shy of 62%.
This number, give or take 1%, has been my win rate for years.
The large losing days were/are what prevented my equity to grow.

I don't think that there is any win rate that is a correct number.

Gabe

If you drop the losing days and retain the winning days, then you're right.

But you still don't get it.

On the other hand, you're not the only one reading this thread, and perhaps others will benefit, either now or in the future.

 

[Rashid_G.]

What a thread this has become!....

My questions...

Without those big down days Gabe's method was making money right?? Why else are we trading? No matter what anyone thinks of the individual entries, it was making money.... for long enough? I don't know.

Therefore what is Gabe looking for here??? If those down days don't dog him from time to time this would be a VERY different journal, if at all.

So question to Gabe, you have a historical 62% win rate without the bad days, are you looking to improve the 62% (could be chasing something elusive) or eliminate the down days?

 

[Gabe2004]

are you looking to improve the 62% (could be chasing something elusive) or eliminate the down days?

All I want is to eliminate those crazy days.
I don't think that having no losing days is realistic.
62% is good enough for me.

Gabe

 

[Rashid_G.]

In that case what lights the fuse on those days that makes a "normal" losing day become a monster? Maybe describing such a day is a good place to start.