Poll: Are you a system trader or a discretionary trader?
- Thread starter WDGann
- Start date
Gann,
What does it make me if my entries are discretionary and my exits are mechanical?
Am I------> 1/2 "discretionary" / 1/2 "system"
I suppose I am like that kid on AI (half mechanical and half human)
What does it make me if my entries are discretionary and my exits are mechanical?
Am I------> 1/2 "discretionary" / 1/2 "system"
I suppose I am like that kid on AI (half mechanical and half human)
I totaly agree with you -- this is a game that can only be played not beaten. Everybody has their own scorecards
WDGann,
I'm intrigued by your hybrid system and wonder what parts of it are mechanical and what parts are discretionary. Do you use intuition to tweak the parameters of a mechanical system? Or does the mechanical system provide overall trading recommendations that you use discretion to fine-tune and execute? I'm just wondering which style operates in which timeframe (e.g., mechanical in the long-term & discretionary in the short-term or vice versa)?
Like you, I do not care who beats who. Moreover, I think that that question is harder to answer than some might think, even if we could agree on a scorecard. It is too easy for some traders of all stripes to be successful (even wildly successful) in the short-term and then blow-up in the long-term. We can't answer the question until we collect everyone's 20-year audited results from actual trading, LOL!
The question of who beats who is hard for two reasons. First, its too easy to create a trading system or discretionary style that is prone to rare, but nasty drawdowns or catastrophes even while it tends to generate substantial profits. Second, its too easy for the market to change -- leaving the system trader with a lot of redevelopment to do and the discretionary trader with a lot of relearning to do (I'm not sure whether redevelopment or relearning is harder???).
The type of trading is less important than how the trader deals with the rare events and the inexorable changes that roil the markets.
Thanx for staring this thread,
Traden4Alpha
I'm intrigued by your hybrid system and wonder what parts of it are mechanical and what parts are discretionary. Do you use intuition to tweak the parameters of a mechanical system? Or does the mechanical system provide overall trading recommendations that you use discretion to fine-tune and execute? I'm just wondering which style operates in which timeframe (e.g., mechanical in the long-term & discretionary in the short-term or vice versa)?
Like you, I do not care who beats who. Moreover, I think that that question is harder to answer than some might think, even if we could agree on a scorecard. It is too easy for some traders of all stripes to be successful (even wildly successful) in the short-term and then blow-up in the long-term. We can't answer the question until we collect everyone's 20-year audited results from actual trading, LOL!
The question of who beats who is hard for two reasons. First, its too easy to create a trading system or discretionary style that is prone to rare, but nasty drawdowns or catastrophes even while it tends to generate substantial profits. Second, its too easy for the market to change -- leaving the system trader with a lot of redevelopment to do and the discretionary trader with a lot of relearning to do (I'm not sure whether redevelopment or relearning is harder???).
The type of trading is less important than how the trader deals with the rare events and the inexorable changes that roil the markets.
Thanx for staring this thread,
Traden4Alpha
I am a hybrid trader - I have some very good quant models but I don't place all my faith in them. My belief is that most systems will fail in the long run. Using my discretion, I trade around the model's signals, which improves consistency and reduces drawdowns.
I will often get a signal and will say to myself "there is no way this one's going to work", which has saved lots of $ by not taking these trades, or waiting for a 'better' price.
Quant models can only code very simple market dynamics, whereas a good discretionary trader can incorporate alot of other factors into the analysis (how is market reacting to news, current sentiment, economic figures, etc). It makes sense to me to combine the two.
I will often get a signal and will say to myself "there is no way this one's going to work", which has saved lots of $ by not taking these trades, or waiting for a 'better' price.
Quant models can only code very simple market dynamics, whereas a good discretionary trader can incorporate alot of other factors into the analysis (how is market reacting to news, current sentiment, economic figures, etc). It makes sense to me to combine the two.
zentrader,
Thanx for sharing. Just to make sure I understand you. It sounds like discretionary "logic" is a post-filter on your mechanical trading system. The mechanical system provides a tentative entry signal, and the discretionary part makes a go/wait/no-go decision.
Do you always paper trade the straight mechanical system to evaluate the value-added or risk-reduced from the discretionary process?
I agree with you about the limitations of simple rules and 100% mechanical. To me, the limitations occur in two time domains. First, a 100% mechanical system may not track the macro-economic picture very well (e.g., foreign capital inflows/outflows, interest rate policy changes, GDP statistics, etc.). Second, a 100% mechanical system may not track the real-time picture very well (e.g., read the tape, interpret the order book, detect the ebb and flow or order imbalances, etc.). From your posting, am I right that macro-economic sentiment (rather than the real-time tape) issues drives your discretionary bias???
Wish you happy hybrid trading,
Traden4Alpha
Alpha,
I never trade against the signal, but will often trade in and out several times in it's direction over a week or so. Most of the time I will exit before the system tell's me so, so I don't follow the model that closely.
I keep track of how my interfering compares to trading the model with no over-riding, and have found that I make less profits but with much less drawdown, so my sharpe ratio is much higher.
Your are right about macro sentiment driving my discretion. My analysis of news/sentiment has been around 55%-60% accurate in forecasting 1-2 day ahead change in the market indexes. For example, pre-market today I was bearish US stocks, because of the general negative sentiment, the fact that we had a 2 day bounce leading in to a weekend, and poor individual company news released after the close. My system also said short so I was comfortable taking the signal.
I don't attempt to read the tape because my timeframe is 1 day or longer.
I never trade against the signal, but will often trade in and out several times in it's direction over a week or so. Most of the time I will exit before the system tell's me so, so I don't follow the model that closely.
I keep track of how my interfering compares to trading the model with no over-riding, and have found that I make less profits but with much less drawdown, so my sharpe ratio is much higher.
Your are right about macro sentiment driving my discretion. My analysis of news/sentiment has been around 55%-60% accurate in forecasting 1-2 day ahead change in the market indexes. For example, pre-market today I was bearish US stocks, because of the general negative sentiment, the fact that we had a 2 day bounce leading in to a weekend, and poor individual company news released after the close. My system also said short so I was comfortable taking the signal.
I don't attempt to read the tape because my timeframe is 1 day or longer.