- Joined
- 13 February 2006
- Posts
- 5,286
- Reactions
- 12,185
Why Do Trading Strategies Stop Working & Why Do They Fail?
Knowing when a strategy stops working is of the utmost importance and it should be to all traders. Monitoring the win/loss ratio, average win/loss, drawdown, and Sharpe ratio sounds well and good but it's not the complete picture I'm looking for.
When a strategy is producing subpar performance, I review it to evaluate the changes in performance metrics. Doing so allows me to make minor adjustments to the trading plan. Before there are negative comments just bear in mind that without ongoing improvements we would still be driving around in a "Ford Model T" motor vehicle. Mechanical system trading is simply an ongoing process of refinement.
Skate.
This is probably the central issue with mechanical trading. Is the system simply in drawdown or broken?
Surely this goes back to an analysis of 'edge'? Assuming you (the trader) understands exactly his edge, the decision as to a drawdown or broken system should be relatively straightforward.
An example: I believe 1 pattern @peter2 trades is a HVBB setup (I may be mistaken in calling this chart pattern an edge, but any number of traders designate a chart pattern as an edge). Look at his examples (charts). Now the trader must examine are HVBB setups still a valid edge (this is simply assuming the results are below backtested results. I have no idea. This is simply a hypothetical edge)? The answer should be a yes or no.
Is not 'refinement' optimisation to prevailing market conditions?
Would not refinement rely on a very small data sample, given that this process could be somewhat ongoing?
Which brings me to my definition of what constitutes an 'edge': an edge must be an immutable, intrinsic reality of the market. If that can be described in the affirmative, then it can only ever be a drawdown, it can never be broken
jog on
duc
Let's talk about the SAP Strategy
This trading strategy I believe has a distinct edge based on a unique "percentage up" buy condition that needs the market to demonstrate a particular level of strength before buy signals are generated. Furthermore, the strategy includes a number of exit conditions that serve to protect earnings and minimise losses.
Exits should be predefined before entering the trade and if you are 100% mechanical then you just execute your plan. If you decide not to execute a predefined exit then you must be clear on the reason for this which may be based on new information that you didn't have when entering the trade. The reason cannot be HOPE.Every trader must possess the essential ability to know when to exit a position. Knowing when to sell can be a difficult decision for any trader, but it is a necessary ability to have in order to control risk and protect your portfolio from large losses.
Exits should be predefined before entering the trade and if you are 100% mechanical then you just execute your plan. If you decide not to execute a predefined exit then you must be clear on the reason for this which may be based on new information that you didn't have when entering the trade. The reason cannot be HOPE.
May i beg to differ here Mr Le Duc on the:This is probably the central issue with mechanical trading. Is the system simply in drawdown or broken?
Surely this goes back to an analysis of 'edge'? Assuming you (the trader) understands exactly his edge, the decision as to a drawdown or broken system should be relatively straightforward.
An example: I believe 1 pattern @peter2 trades is a HVBB setup (I may be mistaken in calling this chart pattern an edge, but any number of traders designate a chart pattern as an edge). Look at his examples (charts). Now the trader must examine are HVBB setups still a valid edge (this is simply assuming the results are below backtested results. I have no idea. This is simply a hypothetical edge)? The answer should be a yes or no.
Is not 'refinement' optimisation to prevailing market conditions?
Would not refinement rely on a very small data sample, given that this process could be somewhat ongoing?
Which brings me to my definition of what constitutes an 'edge': an edge must be an immutable, intrinsic reality of the market. If that can be described in the affirmative, then it can only ever be a drawdown, it can never be broken
jog on
duc
May i beg to differ here Mr Le Duc on the:
an edge must be an immutable, intrinsic reality of the market.
Your edge might be temporary ,so immutable? I do not think so:
You have first knowledge thru technology for example of market movement and so can be first to react, that is a definite edge and the way some darkpool operators are making a mozza, all that can dissappear with a new technology or someone building a data center a bit nearer from an exchange...
and same goes with mechanical trading, you discover a new "recipe" working well, you have an edge, you make money, the strategy spreads, more people are on it and trigger same or similar orders, you miss your buy, sell lower..edge is gone
My edge is being able to manage a batch of trades in such a way that creates an overall profit.
EDGE
The way I see it is, an edge should be as unchangeable as possible. An edge that is unchangeable at the current technological period could be referred to as a real edge but there are various edges of varying life spans and I suppose they could be classified as having a sliding scale of strength depending on how permanent they may be.
I'm really in the middle, I think there can be real edges, strong edges and weak edges, by this I mean a varying strength of an edge.While @qldfrog and I hold one view, you and @ducati916 lean toward the opposite perspective.
May i beg to differ here Mr Le Duc on the:
an edge must be an immutable, intrinsic reality of the market.
Your edge might be temporary ,so immutable? I do not think so:
You have first knowledge thru technology for example of market movement and so can be first to react, that is a definite edge and the way some darkpool operators are making a mozza, all that can dissappear with a new technology or someone building a data center a bit nearer from an exchange...
and same goes with mechanical trading, you discover a new "recipe" working well, you have an edge, you make money, the strategy spreads, more people are on it and trigger same or similar orders, you miss your buy, sell lower..edge is gone
We use cookies and similar technologies for the following purposes:
Do you accept cookies and these technologies?
We use cookies and similar technologies for the following purposes:
Do you accept cookies and these technologies?