Backtesting and Strategy Validation - How Much is Enough?

As soon as an algo has optimisable inputs you are in trouble.

The best algos depend on price action entries.

The problem is compounded with exits. As soon as you use fixed stops and targets you have the issues of changes in volatility to contend with. The fall back position is to use targets that can surely be met 70% of the time in the market conditions that your algo operates in. Stops need to be related to market action and not related to recent swing high and low points necessarily or to a fixed dollar amount.

Yes, some optimising is necessary but picking which optimised inputs to use is important. Picking the almost worst one is better than picking the best but the answer is to look at market action.

What I teach my students is that you cannot expect an algo to work in all conditions all the time. Testing times of day and having the algo trade with different parameters at different times of the day. For example, I run the algo in at least 3 differnt modes: open, middle and closing. There are conditions that I also stop it trading too.

ElectronicLocal
 
The basis of my trading is trading the pullbacks in an identified trend of one chart.

If you look at my blog (google my name), you'll see that I use 2 EMAs (33 & 99) as the basis of my trading. When these go horizontal and the bigCCI I use to judge pullbacks wanders about the zero line, I stop the algo. I also have an algo that identifies the change in trend and I turn the algo on to catch the first pullbacks in that trend.
 
Quote from electroniclocal:

I also have an algo that identifies the change in trend and I turn the algo on to catch the first pullbacks in that trend.
More...

is it automated?

what do you mean by turn the algo on and stop the algo?
 
Have a look at the video on my blog and you'll see what I mean. There is a button on the top left of the chart that allows me, with one click, to turn the automation on and off when conditions require it. Entries and exits are fully automatic but I don't stay passive. I get the instant trade recognition and instant trade execution that I want but I control when the algo runs. I can also move the stops and targets of any particular trade to meet the context's requirements.

The days of just turning an algo on and living in hope are gone. A trader needs to be active in creating an algo that relies on market activity, not blind indicators that are curve fitted, and he needs to be proactive in deciding when the algo should run. It does the hard part of trade recognition and instant trade execution but the trader identifies context.
 
Quote from electroniclocal:


The days of just turning an algo on and living in hope are gone. A trader needs to be active in creating an algo that relies on market activity, not blind indicators that are curve fitted, and he needs to be proactive in deciding when the algo should run. It does the hard part of trade recognition and instant trade execution but the trader identifies context.
More...

I agree.But what so complicated in the algo then?I mean i have it built -in in Ninja Trader strategy for e.g.All i need to do is to set up a target on the assumption of the context and hit the market order.Repeat it in a couple of minutes.

I see it works well with such dynamic instruments as on your chart.
 
That's great, but by using the algos to enter trades I can trade several markets at the same time without missing anything. Also, I have students who take too long to recognise what is happening. The algo does that part of the job without analysis paralysis.

An experienced consistently profitable discretionary trader, under say, 45 years old, trading just one market off one chart will usually out trade an algo. But I'm 65 and I day trade 3 or 4 markets at a time in reasonable size, scaling out of positions, finds the algo to be my holy grail. Its quicker than I am and can pay attention to a number of markets.
 
Quote from electroniclocal:

That's great, but by using the algos to enter trades I can trade several markets at the same time without missing anything. Also, I have students who take too long to recognise what is happening. The algo does that part of the job without analysis paralysis.

An experienced consistently profitable discretionary trader, under say, 45 years old, trading just one market off one chart will usually out trade an algo. But I'm 65 and I day trade 3 or 4 markets at a time in reasonable size, scaling out of positions, finds the algo to be my holy grail. Its quicker than I am and can pay attention to a number of markets.
More...

Not sure of the reason why i should trade several markets at a time.But you are talking about those little things that really works.

And if i asked you to write some algos for me,how much it would cost?
 
Quote from SnakeEYE:

Not sure of the reason why i should trade several markets at a time.But you are talking about those little things that really works.

And if i asked you to write some algos for me,how much it would cost?
More...

(sorry,i have to go now,i`ll read your reply later)
 
I don't do that. It's not in your best interest. I'll teach you how to create your own and I'll show you mine.

You need to "own" your trades and I strongly believe that unless you intimately know and understand the trading logic behind and algo, you cannot have the discipline to trade it and won't know how to manage it. I really didn't post to get biz but I have strong views about the rubbish being sold to people when what really should be taking place is people learning how to create them.

Creating an algo that can trade profitably is a LOT easier than trading as a discretionary trader. You only have to be clever infrequently, not every day.
 
You must log in or register to reply here.
Back
Top