RIP Mechanical Trading Systems?

[Shanghai]

One of the conclusions that I came to a long time ago was that it is highly unlikely that any purely mechanical trading systems can be significantly profitable. There is just too much computing power dedicated to the markets so that any technical or price edge that consistently outperform the markets will soon be discovered and hence be lost. Will also apply of course to fundamental systems based on PE ratios, earnings etc.

The most successful traders I know use simple rule based systems but they use a discretionary element to decide when those systems should be employed, and that's the strategy that I have used for several years and have no intention of changing. So is it necessary to have a discretionary element in your trading in order to make decent money? I definately think so.

 

[the hare]

I definatekly know not !

 

[Hotch]

I disagree, a discretionary rule is just one that you aren't able to define, it doesn't mean it's not a rule, the "computing power" you describe can discover your discretionary rules just as easy as it does mechanical ones.

As you know people who use discretionary methods profitably, I know people who use mechanical methods profitable, most things work in the market if you do em right.

 

[Shanghai]

I disagree, a discretionary rule is just one that you aren't able to define, it doesn't mean it's not a rule, the "computing power" you describe can discover your discretionary rules just as easy as it does mechanical ones.

Not sure that I agree with that point. I generally trade pullbacks to trends but my discretionary element will only considering entering trades when the pulback is orderly for axample. Its very difficult to define what "orderly" is but if I look at a chart I can decide instantly.

Understanding the overall market conditions is an important discretionary part of my trading too. A long trade that I would have entered 3 months ago I won't do today as I think most of the risk to the market is on the downside at the moment.

 

[meanreversion]

One of the conclusions that I came to a long time ago was that it is highly unlikely that any purely mechanical trading systems can be significantly profitable.

At the risk of sounding pedantic, can you define "significant"?

Also, is this not something you can discern from backtests?

 

[meanreversion]

Understanding the overall market conditions is an important discretionary part of my trading too. A long trade that I would have entered 3 months ago I won't do today as I think most of the risk to the market is on the downside at the moment.

I also used to exercise discretion on signals, and kept a record of skipped signals and resultant P/L.. conclusion - don't skip signals.

 

[Rossini]

I also used to exercise discretion on signals, and kept a record of skipped signals and resultant P/L.. conclusion - don't skip signals.

True that.....

 

[Shanghai]

At the risk of sounding pedantic, can you define "significant"?

Also, is this not something you can discern from backtests?

Yes backtesting should be able to give good reliable figures if enough suitable data is available. That would be one major advantage of mechanical systems.

By "significant" I'm probably talking upwards of around 20% per year consistently.

If people are doing well using pure mechanical systems I would be interested to know. But to me it would seem to be a tough job to develop a 100% mechanical backtestable system to return over 50% a year. Anyone that had such a system that gave those returns every year could have so much money offered to them that they would be turning it away.

I guess I'm just curious as to how you could develop a real edge from a pure mechanical stand point given the sheer amount of resources out there that are trying to do the same thing.

 

[Shanghai]

I also used to exercise discretion on signals, and kept a record of skipped signals and resultant P/L.. conclusion - don't skip signals.

I keep those records too and I did tend to skip too many signals. I have fine tuned it a little now.

 

[meanreversion]

Is there an edge to opening a bar or a restaurant? They already exist, doesn't mean that another one can't make money (apologies for the analogy, but note no animals).

The reason I asked about "significant" is that defining a percentage return without putting a number on the volatility of returns is a bit meaningless. If your mechanical system has a Sharpe of 0.5, then 20% vol would give you a 10% return, whereas you would need to tolerate 100% vol (!) to get 50% return.

With sufficient capital, I believe mechanical systems will work. I know you're fairly well capitalized Shanghai, do you have daily NAV values so you can calculate your daily vol?

As for applying discretion, you've been doing it for 6 months and it's been working, but I'm not sure this is a sufficient sample size.

 

[Depth Trade]

One of the conclusions that I came to a long time ago was that it is highly unlikely that any purely mechanical trading systems can be significantly profitable. There is just too much computing power dedicated to the markets so that any technical or price edge that consistently outperform the markets will soon be discovered and hence be lost. Will also apply of course to fundamental systems based on PE ratios, earnings etc.

The most successful traders I know use simple rule based systems but they use a discretionary element to decide when those systems should be employed, and that's the strategy that I have used for several years and have no intention of changing. So is it necessary to have a discretionary element in your trading in order to make decent money? I definately think so.

I think you are right in your statement.
One needs to understand the fundamentals of the product, then be able to mechanically enter the market.

 

[Shanghai]

As for applying discretion, you've been doing it for 6 months and it's been working, but I'm not sure this is a sufficient sample size.

6 months? I've been using the same basic systems with the discretionary element for around 5 years now.

It would be nice to have a purely mechanical system as I prefer not to think given the choice, but I just can't see myself generating the same returns.

 

[meanreversion]

I think you are right in your statement.
One needs to understand the fundamentals of the product, then be able to mechanically enter the market.

To be honest, I've found some of the best moves have come when I've least expected them, or when I've been looking at the other side.

I'm a little dubious as to the long term benefit of fundamental overlay, although I realise this has become a more popular way of looking at things.

 

[Black Swan]

One of the conclusions that I came to a long time ago was that it is highly unlikely that any purely mechanical trading systems can be significantly profitable. There is just too much computing power dedicated to the markets so that any technical or price edge that consistently outperform the markets will soon be discovered and hence be lost. Will also apply of course to fundamental systems based on PE ratios, earnings etc.

The most successful traders I know use simple rule based systems but they use a discretionary element to decide when those systems should be employed, and that's the strategy that I have used for several years and have no intention of changing. So is it necessary to have a discretionary element in your trading in order to make decent money? I definately think so.

hmmm...just wondering if you've got a touch nervy this week given the range us swing traders have found ourselves glued in on certain pairs? Entry? Signal based every time. Exit? We'll never figure it out...ever..So yes you can be mechanical with entry and accept (with swings/position) that you'll leave pips in the move/on the table on exit, or be descretionary and *try* to get the best part of the move...

 

[meanreversion]

Understanding the overall market conditions is an important discretionary part of my trading too. A long trade that I would have entered 3 months ago I won't do today as I think most of the risk to the market is on the downside at the moment.

Sorry, I thought from this statement you implied that your style of trading now is different from 3 months ago? I misinterpreted.

 

[Shanghai]

Sorry, I thought from this statement you implied that your style of trading now is different from 3 months ago? I misinterpreted.

No, I just meant that compared with 3 months ago that we now look well overdue for a pullback on the S&P.

 

[meanreversion]

Come on, the S&P has looked ripe for a pullback for the last 10% .. !

 

[Shanghai]

hmmm...just wondering if you've got a touch nervy this week given the range us swing traders have found ourselves glued in on certain pairs? Entry? Signal based every time. Exit? We'll never figure it out...ever..So yes you can be mechanical with entry and accept (with swings/position) that you'll leave pips in the move/on the table on exit, or be descretionary and *try* to get the best part of the move...

Not so much nervy as bored, so trying to think about something to write about. I trade stocks and nothing has been set up properly for ages and I'm reluctant to enter anything when the index is trading at the top of its channel.

 

[Shanghai]

Come on, the S&P has looked ripe for a pullback for the last 10% .. !

True and its getting bloody tedious staying right up between the 2 and 3ATR on my Keltner Channels for so long. I don't like entering trades when its up at that level so I've missed a large part of the move.

 

[x4x]

Market dynamics change - look at the volatility today compared to pre 2008 - there is no way a mechanical system would handle the difference. That's why EA's are mass marketed when they 'fit' the market, then they 'die' as the market enters a new cycle - then they bring out a new EA, so on and so fourth.

If there was a good EA, there would only be one....