Disadvantages to discretionary trading

[Dogpile 10]

we got a good discussion going recently on the 'Death to Discretionary Traders?' thread. in there, we got a lot of support for why discretionary trading is a good thing...

 

how about a discussion on the disadvantages of discretionary trading? there must be good reasons why discretionary is viewed as negative by big institutions.

[ItalianSharp 10]

Although I strongly believe the advantages by far outnumber the disadvantages, here are a few cons to discretionary trading:

 

1. You must be accountable for each of your losses. There is nobody else to blame but yourself when you make a mistake. Hence, ego sometimes is a big enemy to discretionary trading.

 

2. You must always be on top of your game. Your mind must be sharp at all times the same way a professional athlete must be in top shape whenever he's called out to perform. When you are not mentally fit to trade it's usually reflected in your P&L. And like a professional athlete, you must accept defeat and move on.

 

3. Discretionary trading is a lot more stressful than automated trading. You are the one who is in control of everything, from entry to exit. There is nothing automated in discretionary trading.

 

4. If you are sick or not fit to trade, you must take a day off; which means you are not going to make any money because nobody is trading for you.

 

5. If you get on a losing streak, emotions may become a factor and force you to make mistakes. Confidence is fundamental for discretionary traders. The moment you start to second guess yourself you are utterly lost.

 

These are only a few of the several disadvantages to discretionary trading.

 

I think it all comes down to your type of personality and what fits you best. I started off with mechanical trading and used to get very frustrated at my system. I felt I needed to feel more responsible for what's going on with my money and ever since I chose the discretionary path I've not looked back.

 

Discretionary trading is definitely my trading niche. It's a lot more fun and you learn an awful lot more about the markets simply because your focus must be 110% all the times, whereas with mechanical trading you are more focused on abiding by the rules often times overlooking what the market is trying to tell you.

 

IS

[Dogpile 10]

I have settled on this strategy. Comments appreciated...

 

I spend 100% of my time thinking about the markets current technical structure. What is currently happening in real-time? What is happening on the 15-minute timeframe? What is happening on the 60-min timeframe? What about the 1600tick timeframe? Did we just make a momentum high? Did that push up just 'reject' the previous push down? Are we in a coil? Is this ABC down or ABC up? Is this a higher low or a lower high? Is this a bull/bear flag? Is this good trade location for a long? Does trade location favor the pattern I am seeing?

 

I want to try to be flexible on this discretionary part as it plays out. When I get a sense of where it is going, I then wait for a mechanical scalping strategy to confirm it.

 

I have a few mechanical strategies lined up and optimized for profit targets and stop losses... A few short-focused and a few long-focused. All use stops for entries and exits. I figure if they are profitable on their own, I can make them much better.

 

It is up then up to me to figure out where the market is trying to go and 'go-with' the long signals if odds are to upside and 'go-with' short signals if odds are to downside.

 

Is this a common approach? Do any of you guys do this? If you are right on the structure, you make a lot of money. If you are wrong, you stop out. It is then all about what % you are right. when you are wrong, you are wrong and you move on and wait for the next pattern and confirmation from the mechanical strategy. This has really been working for me lately and curious if this is a common approach?

[torero 10]

Unless your mechanical strategy's rules are based on the same assessment as your discretionary strategy, then you're comparing apples and pears. They have two different structures viewed by two types of technical analysis. I've tried this myself trying to read the market my own way as the auto trade is in progress. Most of the time, I take over, I end up disappointed with the results, either cutting short my profits or exit before it has a chance to move into profits.

 

The problem with discretionary strategy is that we cannot see the way we program the systems anymore. You mentioned bear flag, you're used to looking for it and not looking at the stochs (example of your mech system) meticulously for an entry anymore, your judgment is no longer trained to view that way. I've had this problem and I have the charts where the strats are turned on and set them aside and not have it interfered with my discretionary decisions.

[Dogpile 10]

not sure I understand what you mean:

 

<<Unless your mechanical strategy's rules are based on the same assessment as your discretionary strategy, then you're comparing apples and pears. They have two different structures viewed by two types of technical analysis. >>

 

there is one structure but it is being defined by the mechanical rules.

 

for example, I see this pattern over and over again. the market closes low (or high) in its daily range, it will very often make a lower low (higher high) the next day followed by a reversal in the morning session. let's just say this is a tendency. ok, what defines a reversal? well, its pretty ambiguous...

 

you can do an average true range function and if it goes X true ranges off the lowest low, than that is a reversal. now you can test for this mechanically. that is your structure. doesn't mean that if it goes X true ranges off the low that it is a profitable trade, it just means you have defined that to be a reversal and anything less than that is not a reversal. this is a consistent structure. if it doesn't go X true ranges off the low, then it was never a reversal, despite you thinking it was going to do that (your discretionary thinking at the time). so long as you are playing for reasonable profit targets, you should be ok. if it is a reversal, you will make your profit. if it isn't a reversal but flushes in that direction, you may make your profit anyway before the reversal fails. if it starts to reverse but then fails before hitting your profit target, then you lose.

 

my real point is that you are just trying to find an entry/exit technique that is reasonably good statistically -- on a lower timeframe. the money-making will not come from this technique, it will come from your proper identification of the higher timeframe structure. I say this because after you write a simple mechanical strategy, very often you can see tons of trades that are clearly stuck in some range that you can easily discard as you never would have taken these 'signals' (this happened today, Friday). yes, some of your discards will be bad as they will break out of the range on that signal --- but that is similarly true if you are just discretionary trading.

 

just a little confused on what you meant by saying they have 2 different 'structures.'

[torero 10]

Ok. I see what you mean. Basically you're using the aid of the system in identifying your entry. My problem was identifying the structure different from my system. Basically I used it as an entry but ending up trying to exit with a different indicator.

 

If you use it this way, it's fine. You'll have higher percentage of winners with smaller profits, it's ok. So long as you tighten up your stops for the losers, but that will distort your results from the original system. When these failed reversals start kicking in more often, the small winners may not cover the larger losses. I don't think there's anything wrong in system aiding in identifying your entry setup.

 

I have an excel keeping results on the different profit target distances (ie. 2, 4, 6pts for er2) to determine if the optimal exit point. (I know I can do this on TS but I use all discretionary setups myself without auto trades.

[waveslider 10]

Dogpile,

 

I think you have a valid idea. I have an entry method that statistically and logically works. I use discretion, looking at various things to bias myself direction-wise, then use a non-quantitative method of exit that is dictated by the market's reaction to my entry.

 

The point is, you have to have a valid entry that works statistically. Even if just from a risk-reward point of view.

 

As discretionary traders I think this mechanical entry is an immense value for us because it halts us from opening up positions just any old time the market looks good or bad to us.

 

I had a tendency to overtrade, still do. By waiting for the method to play out, it increases my discipline.

 

ws

[Dogpile 10]

<<So long as you tighten up your stops for the losers, but that will distort your results from the original system. When these failed reversals start kicking in more often, the small winners may not cover the larger losses.>>

 

the only 'system' is on the lower timeframe. my stop is based on the lower timeframe -- there are no larger losses possible because the initial stop is not discretionary -- its based on the optimization I have already performed. now if it moves my direction, I will always take 1/3 to 1/2 the position off at least. the other piece will be a discretionary exit but never a stop further away than the original stop -- and usually I tighten the stop to small risk or breakeven depending on the initial push after entry.

[Dogpile 10]

waveslide, we should chat a bit more --- I really like this method and sounds like you may have been down the same road as me before you arrived at this...

 

<<I have an entry method that statistically and logically works. I use discretion, looking at various things to bias myself direction-wise, then use a non-quantitative method of exit that is dictated by the market's reaction to my entry.>>

 

I would love it if you had time to discuss an example of a trade you did with a few points about what made your 'bias' -- the real important part... but also, what your entry method was... I don't mind chatting about this publicly ( in this forum) if you don't... its really not like some giant institution can possibly emulate our strategies...

[Dogpile 10]

another nice thing about having a mechanical short-term strategy is that you have a daily benchmark to outperform. you can see the signals it generated and see which ones you took and which ones you didn't and why... you also see when a mechanical trigger is about to go off and have to time to think, do I want to take this signal or not? does the higher timeframe technical structure favor taking this trade? is the location favorable relative to the support/resistance you have identified? is it a really choice set-up consistent with a full-size position or do you see a few reasons for concern and should only take a smaller position? or is the set-up really pretty balanced with no edge whatsoever??

 

Friday was an example of a difficult day to trade in my opinion... my strategies did not do well on Friday after having perfect signals on thursday. you can see how Friday was not really a good 2-way market. good traders, whether bullish or bearish, had plenty of opportunities to make good money on thursday. I made money on both days but underperfomed my ('dumbed-down') mechanical strategies on Thursday and outperformed on Friday (though I didn't make much on Friday, it was a green day). if I summed the 2 days together, my (single-timeframe) simple mech strategies were basically back to breakeven by the end of Friday...

[darthtrader 10]

really interesting ideas, please don't take this conversation private.

 

sounds like a great way to get the best of all worlds trading wise, i had never really thought of things in this way.

 

dogpile, i take it then that while you will cancel a trade based on the mechanical entry you won't enter without it?

[Dogpile 10]

<<dogpile, i take it then that while you will cancel a trade based on the mechanical entry you won't enter without it?>>

 

right. I could try to figure how to program all the usual support/resistance pivots I follow into the strategy and test it --- but I am really a 'pattern guy' so I prefer to have a really simple entry strategy (Keep It Simple doctrine) and try to think hard about the pattern and the 'location' of the trade rather than stick a bunch of filters into the program and have all these offsetting forces in there potentially screwing it up....

 

I am a believer in finding an indicator that really 'speaks to you' --- something that fits your personality and your style... and most importantly, something that keeps you out of trouble.

[walterw 10]

<<dogpile, i take it then that while you will cancel a trade based on the mechanical entry you won't enter without it?>>

 

right. I could try to figure how to program all the usual support/resistance pivots I follow into the strategy and test it --- but I am really a 'pattern guy' so I prefer to have a really simple entry strategy (Keep It Simple doctrine) and try to think hard about the pattern and the 'location' of the trade rather than stick a bunch of filters into the program and have all these offsetting forces in there potentially screwing it up....

 

I am a believer in finding an indicator that really 'speaks to you' --- something that fits your personality and your style... and most importantly, something that keeps you out of trouble.

 

 

Wise, wise wise....

[waveslider 10]

To me, any entry approach that offers :

 

1. a good number of trades

2. participation in most (all) major moves

3. a good risk/reward profile

 

is valid for use.

 

Stand alone, any single time frame mechanical strategy (that we as individual traders are capable of following) is not likely to be successful in the long term.

 

What brings the odds into the discretionary trader's favor is bias, and I see the most important bias in the higher time frames. Why? Because of what old Newton said, "an object in motion tends to remain in motion". In trading this would mean "a market in a range will tend to remain in a range" or "a trending market will tend to remain trending". Take this information from the higher time frame and bias your entries on the lower time frame.

 

Dogpile, your parabolic system is just as good as any. If you were on friday to see on the higher time frames a range, or a downward trend, and bias your trades to the short side only, I'm sure the approach would have been successful.

 

So this is the double edge sword - back to the topic of the thread. By biasing you are manipulating the results of the mechanical system.

 

The challenge to discretionary traders is to see objectively and recognize things the computer may not. As many have said before, trading is an art more than a science. This is a skill that must be developed and practiced, an understanding must be developed of how markets demonstrate who is in control.

 

I believe there are some very successful systems out there that trade 100% mechanically, but I don't think they will be as robust over the longer term as the skilled human mind.

 

Personally I prefer to use just basic trendline analysis. Something as simple as this is really all you need. The way it leads to profitability is through discipline and consistency. The lack of these 2 traits is the true disadvantage to discretionary trading.

 

ws

[Dogpile 10]

<<Personally I prefer to use just basic trendline analysis. Something as simple as this is really all you need. The way it leads to profitability is through discipline and consistency. The lack of these 2 traits is the true disadvantage to discretionary trading.>>

 

yes, I can see how trendlines would make sense. nice thing about trendlines is that you will never get caught fighting a strong move as the entries will simply not trigger when this is the case.

 

would like to see a recent example of what kind of trendlines you are talking about? can you post a chart with an example of an entry? I would be interested in experimenting with drawn trendlines on YM

[waveslider 10]

I'll start a new thread..