Getting Started Trading Systems Combining Mechanical and Discretionary Trading

Mechanical systems do not work.

Let me rephrase that. Backtested over several years of data, mechanical systems do not show a sufficiently consistent profit and an acceptably low drawdown for the small trader to trade consistently, comfortably, effectively and economically.

To expand on that further, the results of a coded mechanical system might well show periods of excellent profits but, virtually without exception, they will also have draw-downs that most small traders would not be able to tolerate. In addition, most systems tested over several years invariably have considerable periods of losses, usually running over months if not years, plus long runs of losing trades.

If you start trading a mechanical system at the wrong time, you could face months of losses before it starts to turn profitable. Few small traders have the stamina to do that or to sit through more than a handful of consecutive losing trades.

But, you protest, the trading forums, including this one, are full of people using mechanical systems quite successfully. Apparently so, but if you look closely at these posts you will often see a request for someone to program the system so it can be backtested mechanically. None, to my knowledge, have produced consistently profitable results over any meaningful period of time - although if you know of any which have, please point me to them!

So how can some traders claim success with a so-called mechanical system that can often demonstrably be proven not to be successful over the long term and that other traders fail to achieve success with at all?

The answer is that they use discretion. Their own discretion.

If you look at what successful traders do, you'll notice that, almost without exception, they all rely to a degree on their own intuition, feel and instinct - discretion. They may use indicators, chart patterns, Fibonacci levels - all manner of Technical Analysis tools - but the setups and trades are filtered through their own perception of what's happening in the market. These are the traders who consistently make money.

Sure, they have bad days, bad weeks, and they have drawdowns but they have one thing that a mechanical system does not have - a brain. And they use it! All traders have a brain even if they don't all make the best use of it! Although the human brain may not be able to match a computer for sheer processing power it's far better at pattern recognition, fuzzy and deductive logic, at looking at a wide variety of disparate inputs and signals and pulling all these together to form a conclusion.

Mechanical systems with discretion

In my own trading I have found that the way to success in trading lies in combining a set of trading rules, ie a mechanical system with discretion?

This can be incorporated into many trading systems - in essence, use a system or methodology you feel comfortable with but don't automatically take every trade it throws at you. Think about the chart, the price, the setup;
ask yourself how you feel about the trade. Does it look right?

Here's an example. Let's say one of your setups is the RSI hooking down after it's reached an overbought level. Let's say on one occasion that it hooks when it's 1 below the actual overbought level. Do you accept that as a Go signal or not?

If you were trading purely 100% mechanically, you would not. However, if, to you, the other signals look "in good order" then you might "give yourself permission" to take the trade. In other words, don't restrict yourself by waiting for all the Is to be dotted and Ts to be crossed. Take in the wider picture and make a decision accordingly.

Another example - let's say all your setups are firing and you're ready to put on a trade. Before pulling the trigger, take a moment to look at the bigger picture. Okay, the RSI may be well into overbought territory and starting to hook down but, wow, look at that rising trend. The RSI is quite happy to stay overbought or oversold while the market continues in its original direction so does the price look likely to continue or to reverse? Use your judgement. If you're not sure, wait for confirmation that the market has reversed - a chart pattern, a break through a support level, something that adds credence to the system's signal.

Now you're not going to make the right call all of the time - but then neither does any system or trader. The object is to use your skill, judgement and experience to make better calls than a purely mechanical system so you get the best of both worlds.

Now, some of you may be thinking - that's all very well, but I simply can't trade intuitively. That's okay. Some people seem to have a natural instinct for it but most of us must learn it. It is a skill and it will improve with experience. As you learn, you will be looking at and evaluating the charts. You've probably heard this before but there is no substitute for screen time - watch the markets, see how they move and how they react to various situations. Watch diligently and you will learn by osmosis.

Starting out

The first thing you need is a system or method. The Net and trading forums are full of systems. Select one that seems logical and reasonable, one that makes sense to you. Remember, the object is not necessarily to trade every signal in every situation but to look at the bigger picture and make an informed decision about whether or not the setup is valid and is likely to make a good trade.

Many beginners start off by looking for a mechanical system hoping or expecting that it will automatically put them on the road to riches. After a few months' experience, often to the detriment of their trading account, they realise that it's not as simple as that.

While beginners can certainly start by looking at mechanical systems, the acquisition of the necessary skills to interpret charts will take a little time. Remember this - if it was easy, everyone would be doing it. It isn't and they're not. But like most things in life, it is a skill that can be learned and one that will improve with experience.

If you are a newcomer to trading, may I suggest you spend time learning the basics of trading, money and risk
management, order placement and a spot of trading psychology? This is all freely available on the Net, particularly in trading sites and forums such as this.

For more about the inconsistency of purely mechanical systems and the necessity to learn how to actually trade, take a look at Malcolm Robinson's excellent article, The Evolution of a Trader.

The road to a combined system

If you're not sure that you can develop your own discretion, you can build up to it by exercising your judgement in a more, shall we say, practical way. For example, before jumping into a trade, see if the price is near a support or resistance level, a fib level, a recent high or low. If it is, this might suggest that although the basic premise behind the trade is sound, the price may not move very far creating a poor risk/reward ratio.

Check if any news is due that might affect the market. Forex, for example, often moves significantly - and unpredictably - on the release of certain news reports so you need to take care when trading around those times.

You can also use the "look before you trade" principle if you work with several indicators. A purely technical trader might want all his or her indicators to be firing at the same time before taking a trade. If you add a discretionary element to your trading you don't have to do this. You can weigh up the importance of each indicator and even if one is not firing, take the trade if everything else looks right. Or, if one is lagging behind the others but it looks as though it might fire in the next bar or two, you might use your judgement to step in front of the signal and enter early.

You can use your discretion in any situation and apply as much or as little of it as you like. To begin, when your system signals a trade, you might use discretion simply to check the price against support and resistance levels or trendlines. As you gain experience you might take fib levels or pivot points into consideration. As a setup approaches, you might search for candlestick patterns such as dojis or chart patterns such as double tops and bottoms that might confirm the trade or suggest caution.

In essence, what you are doing is using the power of your brain - its reasoning, deductive and intuitive abilities - to adapt your trading method to changing markets. With experience, you will do this naturally, subconsciously. We allow and compensate for narrow-range days, poor volatility, erratic price movement - these and a whole range of other factors that make the market move as it does, factors that standard indicators and technical analysis can't fully take into consideration when presenting their signals.

We can do this - purely mechanical systems can't. Self-adapting and self-optimising systems attempt to do it but they are not consistently successful, and there is a whole raft of Artificial Intelligence, Neural Network and Genetic Algorithm software that aim to analyse markets to the Nth degree to create successful trades but none can match the mind of a trader well-tuned to the market.

Caveats

Like any trading system, it's possible to lose money with using a combined mechanical/discretionary system just as easily as you can make it. To make money, your judgement must be correct and this will only come with practise and experience. However, you can be sure that blindly following a mechanical system will not, in the long term, make you profitable.

Also remember - the entry decision is only part of the trading process - and some would say only a small part at that. However, without a method you have no trades at all so the system remains an essential part of the plan.

By combining discretion to your trading rules you will begin to develop your own trading style. Your decisions and judgement calls will be based on aspects of trading that appeal to you and that you are comfortable with. It can be a self-reinforcing process. As you make successful decisions, you become more confident about them which, in turn, makes you more comfortable with using your discretion.

If you have any questions or comments, please feel free to contact me.
 
Last edited by a moderator:
What normally happens when you apply discretion to a mechanical system is that you end up
not taking the best trades. The best trades are from obvious and often tend to come out of
nowhere.

The article mentions that drawdowns that are the biggest problem with mechanical systems,
a disciplined trader will not have problems taking drawdowns, if you are risking 1% of account
per trade, drawdowns in excess of 25% are very unlikely. And if you are trading multiple
markets and a combination of non correlated systems (trend/counter trend combined) these
drawdowns wont last very long.

A small trader with 10K account wanting to make say 5K a year, should be willing to take a 2.5K
drawdown.

If you cant handle drawdowns you shouldnt be trading.

This is not to say that living through drawdowns is easy, it isnt, but you just have to do it
if you want to be a succesfull trader.
 
Last edited:
Drawdowns can be managed and minimised by running a portfolio of trading systems.
 
Drawdowns are inevitable!
But would you want more drawdowns than necessary?

Is there a way to avoid drawdowns when possible? Of course there is. You need to apply logic and research to your trades.

For example, if your mechanical system gave you a long signal, but price was sitting a few pips below a major resistance level, would you take the trade at that time or would you wait a few pips before you placed the trade?

If you decide to take the trade because you are a hard core mechanical follower, than this will result in more drawdowns than necessary.

If you decide to wait a few pips, than you have just used a part of your brain to apply discretion (Mechanical Discretion to be exact).
This is what the artcle is referring to. Get all the facts before you place every trade.

How often do you find taking a trade gives you a profit of 5 pips before it turns around and hits your 20 pips stop loss? These types of trades can be avoided at times and therefore decreasing the number of losses and drawdowns.

JH
 
If the resistance level is that important one could include it in the mechanical system. If not and if the system has been build correctly it really shouldn't matter.

I don't think there is any problem about using discretion as long as it doesn't mean you end up not taking trades. One way to do it is to have a discretion window on a trade. The system says buy now, but you could give yourself a time window to do the buying. One can then performance test yourself against the system.

DonaldDuke is right about Drawdowns. The best trades are normally the ones that you really don't want to take as they just seem wrong. Sounds like the systems you've looked at Jetheat are fitted to market character so they come and go (eg hi volatility = it works, lo volatility = it doesn't).
 
discretions are probably better applied to exits, rather than entries..

missing out on good gains is psychologically tough, and you could end up with the trading equivalent of "the yips".
 
I know exactly where you're coming from JH - but I do have concerns about the impact your article may have on less experienced traders.

There can be no 'real' hybrid mechanical/discretionary trading system - they are a contradiction in terms. What you are explaining as it exists within your trading approach is (presumably) born from a great deal of personal experience with your chosen market area(s).

This building of discretionary capability you allude to is another name (I believe) for Intuition. A much maligned word in trading circles, but if you consider Intuition is nothing more the Unconscious processing of information events. And that these events you are only able to process Unconsciously because of long exposure and experience to and with them within conscious awareness - it makes a lot more sense.

Like the 'Intuition' the car door hasn't shut properly. How many times have you 'heard' it shut properly? And how many times have you 'felt' just how much effort was needed to close it properly? I may be dragging the point - I hope not. You 'know' the door hasn't shut because it hasn't 'matched' one or more of the criteria which (by now) you Unconsciously process each and every time you carry out this routine event successfully.

So for those traders who have put in the hours and have moved into the realms of Unconscious awareness and trading has become a 'routine event' - fine. But my concern is that many newbie traders may well decide they already 'know' enough to circumvent a profitable system and method - to their ultimate detriment both financially, psychologically and experientially.

A large number of hours expended working with a mechanical system may well lead to Intuitive insights and discretionary capability. But these insights 'arrive' in and of themselves and are unmistakable when they do. My own belief (and it is just a belief) is that this is a natural process very much determined by the unique environmental, mental and physical characteristics of the individual. To suggest or propose the active acquisition of such skills may be worse than self-defeating.

I look forward to other members comments and experiences using your approach.
 
Last edited:
This is my point exactly.
What I'm trying to convey is, new-intermediate traders need to learn a bit more about how, why when etc.. before they start to apply their own discretion to a mechanical system.

Of course, applying your own discretion to a system has to be correct otherwise there is no point.

I believe discretion can be taught to new-intermediate traders and I also think it's importanat for these traders to get on the case as soon as they can otherwise blindly following many mechanical systems can be dangerous.

JH
 
Drawdowns are inevitable! But would you want more drawdowns than necessary? Is there a way to avoid drawdowns when possible? - JetHeat.

Hi JetHeat
I agree that drawdowns are inevitable, and I certainly would not want more drawdowns than necessary. One way of trying to avoid drawdowns being a huge issue is to ensure that in the design, backtesting, evaluation and maintenance of an MTS (with tradestation or metastock etc.), you develop an MTS that does not have huge drawdowns!

Although i see merits in your logic of checking to see that there are no factors present that may influence the effectiveness of a placed trade that your MTS may have missed, for me, the goal of using an MTS is to remove the discretionary aspect of trading, making trading as objective as possible, rather than subjective. Trading with the belief that the historically profitable MTS will make an overall profit.

I would not want to second guess my MTS, as this would undermine my purpose of using the MTS. I would also then find it infuriating to miss big moves that my MTS had identified, but which i had not traded. I would rather be able to blame the MTS for any losses rather than myself!

Obviously I would use discretion as to whether to execute my mechanical trade entry and exit signals around major news releases, but at all other times would otherwise intend to trade every trade entry and exit signal.

If, as i progress, I see occurrences within the chart that could be taken into account within my MTS and it's coding - which will improve the MTS - I can modify the MTS accordingly.

Thanks

jtrader.
 
Last edited:
mmm, I wouldn't think there are many mechanical systems that can be operated without some discretionary element being involved somewhere along the line twixt entry and exit. If you had a rule book to cover every conceivable eventuality it'd probably be so thick that you'd spend all your time reading it and never get around to trading.

For those who do claim to have such a system, why are you sitting in front of your screen instead of letting your computer do it? I suspect it's because you need to be there to make the "discretionary" decisions that your computer can't make.

Aside from all that, discretion comes into play a lot earlier. Give a hundred people a workable system and then look to see what they are doing after a couple of months. How many of them do you reckon will have exercised their discretion and changed the rules a bit :LOL:

good trading (discretionary or otherwise)

jon
 
One of the biggest advantages of pure mechanical trading is to avoid human emotions like
greed,fear,boredom,vengance,self blame, regret etc

Trading with a mechanical system and using stops for entry and exits helps to avoid these emotions (atleast at the individual trade level)

When you bring discretion into the equation you open yourself to these emotions.

For me personally it would be a counter productive. For other people it may be a help rather than a hindrance.
 
Jet Heat, Great article and I agree entirely a purely mechanical system can never work in all situations at all times. If it did then one would be a billionaire within months. Besides everyone would use it and it would then stop woking. Unfortunately purely mechanical systems only work in the past or when paper trading. Look at al the mechanical systems for sale why so many? In theory you only ever need one mechanical system "The one that works". If I ever developed such a system I certainly would not sell it to anyone else I would just make money out of it.
 
Discretion

FetteredChinos said:
discretions are probably better applied to exits, rather than entries..

missing out on good gains is psychologically tough, and you could end up with the trading equivalent of "the yips".

My personal opinion is the complete opposite to this. I apply plenty of discretion to entering a trade but very little to exiting (especially a losing) a trade.
 
"Unfortunately purely mechanical systems only work in the past or when paper trading."

What basis is this statement founded on?
 
In the book Tape reading for the 21st Century by Clif Droke, Linda Bradford Rascke, a wellknown trader who traded also on the exchange (that is my understanding) is quoted as saying "I have known hundreds of professional traders throughout my career. I don't want to disappoint you but I know of only two who were able to make a steady living for themselves trading with a mechanical system."

Do any of you guys know anyone who has been able to make a steady living trading a mechanical system and if so what system?

PT
 
pttrader said:
"I have known hundreds of professional traders throughout my career. I don't want to disappoint you but I know of only two who were able to make a steady living for themselves trading with a mechanical system."
Two things to consider here.

1. Professional traders have limitations and constraints which are not normally applicable to the vast majority of individual private traders. Their failure/success rates can't be compared with an unlike sample.

2. How many individual private traders are ever interviewed for these types of books? How would the writer/researcher ever find out about them?

Some mechanical systems work.

Some discretionary trading systems work.

I've never known of a mechanical system that has benefited long-term from any influence on it of a discretionary nature.

You are unlikely to get a response to your final question of these boards. And I'm a little surprised you asked.

There are a number of system vendors on these boards who may well be contacting you privately, but they're in for a bit of a surprise I imagine.
 
Mechanical Discretion

I agree,

You cannot use a mechanical system entirely on its own.
Sometimes, you have to use discretion to avoid the obvious traps.

The "mechanist" would argue that you could miss out on big moves when you don't take a trade.

Generally, you end up losing more on a mechanical system if you don't use the correct discretion. There may be a few mechanical systems which produce more wins than losses but more often than not, pure mechanical systems need human intervention.

I call it "Mechanical Discretion (MD)".

New to Intermediate traders need to know the basics so they may be able to apply discretion to their trades.

In my trading ( and in those who have followed the MD rules) have experienced much better trades simply because they knew how to refuse potential "Death Trap" trades.
 
You cannot use a mechanical system entirely on its own.
Sometimes, you have to use discretion to avoid the obvious traps.


And sometimes the trades that look like obvious traps turn out to be the biggest winners. This happens more often than you might think.

Generally, you end up losing more on a mechanical system if you don't use the correct discretion.

Obviously using the 'correct discretion' will improve a system, however, it's not possible to know whether or not it's correct until after the event. You cannot test discretion like you can test a system, which is the whole point.

There may be a few mechanical systems which produce more wins than losses but more often than not, pure mechanical systems need human intervention.

This seems like an ill-informed generalisation - there are plenty of systems that are profitable, but they will not work 100% of the time. This is where the intervention is required. You must decide to trade a system based on its' performance & then stick with. Review it periodically and then decide whether or not to amend/continue.
Trying to cherry pick which signals to take may mean you miss a few losing trades, but at the expense of missing some of the biggest winners & getting poor entries on all trades whilst you work out if you should take them.

In my trading ( and in those who have followed the MD rules) have experienced much better trades simply because they knew how to refuse potential "Death Trap" trades.

There are no 'Death Trap' trades in a properly designed system that includes appropriate money management rules.

For most traders, using discretion will reduce the performance of a mechanical system and remove all benefits in terms of reducing stress/removing emotions etc. Another of the main reasons for using mechanical systems is that you've already proved to yourself that the system makes more money than you do. If that's the case it seems foolish to ignore it in favour of your proven inferior judgement :!:

Simon
 
Two things to consider here.

1. Professional traders have limitations and constraints which are not normally applicable to the vast majority of individual private traders. Their failure/success rates can't be compared with an unlike sample.What are the limitations and constraints? I would think if anything the professionals have an upper hand on the general public.

2. How many individual private traders are ever interviewed for these types of books? How would the writer/researcher ever find out about them?That is fair. However, it would be interesting for some to come out of the woodwork and show a mechanical system that has worked consistently for them say over a period of 3 years. Not historical testing on past data but actual trades taken. I am not saying can't be done but just saying it would be very interesting to see a pure mechanical system that works well for 3 years straight.

Some mechanical systems work. Perhaps, I would like to see one with actual trades actually taken.

Some discretionary trading systems work. That is a fair. Many don't.


You are unlikely to get a response to your final question of these boards. And I'm a little surprised you asked.

There are a number of system vendors on these boards who may well be contacting you privately, but they're in for a bit of a surprise I imagine.I would have to see brokers statements not historical testing nor real time testing but proof of actual trades made over a significant time period then I might be interested. Otherwise, they might be in for a surprise as you say! By the way the reason I asked is I am of the belief that it is unlikely that mechanical systems over the long haul are a viable method of trading. It would be good to see one that has proven viable. I am certainly open to change my beliefs about it. The market changes so much that it seems to me judgement calls have to be made to react to the market. I am by far a discretionary trader. There are some mechanics in the system I use but it is not by any means purely mechanical.

PT
 
Last edited:
Mechanical Trading

I think the most publicised and famous mechanical trading system was that used by the turtles, but even they were allowed some discretion for instance whether to use a 10 day or 55 day breakout.
The reason I feel these systems only work in the past is that this is true for any system even with discretion. Looking at historical charts I am a 100% perfect trader because I can spot anything and give good vaild reasons why the share behaved as it did. Unfortunately at the RHS of the chart when there is no future to see I am not so good. With mechanical systems when you look at the figures and sometimes the drawdowns I wonder how many people have the discipline and indeed the pure balls to trade through it. I have no doubt mechanical systems can work in certain markets and certain situations however how does one know when the market and situation changes and therefore when to stop using the system. For instance is there a massive drawdown (and things will turn out OK) or does it simply not work anymore. There are 100s of mechanical systems for sale do they all work? If you had such a perfect system would you really sell it off at a bargain price, I don't think so.
 
Top