Identifying an edge

[Visaria]

See the second post, by the legendary acrary:

http://www.elitetrader.com/vb/showthread.php?s=&threadid=106057

The second paragraph is something you guys should drill into your brains.

 

[erol]

Thanks again guys that's great...

If I am consistently making money I would suspect that I have an edge, but not conclude it, because I realize things change.

However I'm talking mostly about the process of identifying an edge. Maybe I should have asked how "does one go about identifying a trading strategy?" (not trading plan), but it sounds like that is personal as well.

Anyways, this is helpful feedback, thanks again!

EDIT: intradaybill and Visaria, those are helpful threads thanks!

 

[Epic]

Quote from Visaria:

See the second post, by the legendary acrary:

http://www.elitetrader.com/vb/showthread.php?s=&threadid=106057

The second paragraph is something you guys should drill into your brains.

+1

Edge is identified in the ability to take advantage of the non-random events. Much of the price action in the market is indeed random, but the market can best be described as randomly wandering between non-random events. Hence the trending nature of the markets.

TA critics use the inability of TA to consistently predict future price as the basis for their argument against it. They are looking for proof of a TA system that is accurate at least 75%.

Proponents use their ability to accurately predict action even on an occasional basis as proof of its effectiveness. They argue that 75% accuracy is not necessary, and in fact even 50% accuracy is not necessary.

IMO, both are correct. Many tickers are indeed random a majority of the time. It is difficult to develop an edge based on price action for this very reason. There is simply a ton of noise. An edge usually consists of proper execution of trades, which is why you'll generally hear comments regarding the exit as more important than the entry.

Backtesting is a difficult way of identifying an edge mainly because purely price action based edges are difficult to develop. Many times the edge that a traders purports is actually just chance landing them above the mean on a probability curve. This is why so many edges are lost as the market shifts. The supposed edge was simply a chance agreement with medium term market conditions. A true edge transcends market conditions and will allow the trader to outperform peers under all circumstances. Such an edge can be enhanced by a larger bankroll but frequently becomes hindered when scaled significantly.

In the end, to me, an edge must be identified through the failures of a strategy. This happens as the trader employs the strategy over time. At some point the strategy will have been tested over various market conditions and it then becomes verifiable. Fortunately, these days a trader will experience all possible market conditions over the course of maybe a 5 year period.

 

[1a2b3cppp]

Quote from Visaria:

This 'method' doesn't determine that at all. I just need one example to disprove it. I'm sure LTCM's equity was rising every day for literally years before they collapsed.

So you could say LTCM's "edge" ceased to be an edge when they collapsed.

 

[Epic]

Quote from 1a2b3cppp:

So you could say LTCM's "edge" ceased to be an edge when they collapsed.

LTCM's edge ceased to be en edge when they stopped using it. The problem with their edge was the same as with most edges. They do not scale indefinitely. LTCM continued to scale their strategy well past the point that the edge disappeared.

A person would have a tough time making a case that LTCM didn't have an edge when all the facts are considered.

 

[wrbtrader]

Quote from Epic:

...In the end, to me, an edge must be identified through the failures of a strategy. This happens as the trader employs the strategy over time. At some point the strategy will have been tested over various market conditions and it then becomes verifiable. Fortunately, these days a trader will experience all possible market conditions over the course of maybe a 5 year period.

Well said but I would like to add "an edge must be identified through the failures and successes of a strategy".

Mark

 

[Ghost of Cutten]

Quote from erol:

Hello,


I've accumulated a lot of knowledge on how things work, and what to do when certain conditions are in my favour, however I suppose the challenge is knowing when to execute a trade.

I was wondering how you successful traders with an "identifiable edge" found your edge.

Is it a matter of data mining? Identifying patterns and back-testing to check if the validity of your (proposed) strategy?

I guess I'm just asking for a map, or a starting point.

Observing the markets for a long time, and recording the observations and my thoughts in a trading diary. Taking notes on any repeating patterns, then trying to design methods to exploit them while controlling risk. Then, learning from attempts to apply those methods. Continually revising and improving as more experience and observation is accumulated.

Most useful learning was done via thorough post-mortems of mistakes, losses, and missed opportunities.

That's enough to get to moderately profitable. After this - preparation, training, systematisation of methods, and rigorous & disciplined implementation help improve performance beyond mediocre levels to acceptable or superior performance.

 

[Epic]

Quote from wrbtrader:

Well said but I would like to add "an edge must be identified through the failures and successes of a strategy".

Mark

Sure, but through the methodology that I'm referencing, the successes need not be identified individually. Only the failures are important because you don't need to adjust for the successes, just the failures.

1) Backtest to ensure that you aren't about to employ a guaranteed losing strategy.
2) Identify worst case scenario and risk tolerance and size positions accordingly.
3) Trade real money.
4) Keep accurate notes and perform objective analysis.
5) Identify all failures and the reason for failure.
6) Adjust parameters to either avoid failures or limit magnitude of drawdowns from inevitable failures.
7) Employ strategy through all conceivable market conditions.
8) Track Sharpe/Sortino ratios over time.


Once the majority of possible market conditions have been experienced, then the ratios will give you a pretty accurate indication of whether or not you do have an edge.

 

[piezoe]

Quote from Visaria:

This 'method' doesn't determine that at all. I just need one example to disprove it. I'm sure LTCM's equity was rising every day for literally years before they collapsed.

Well, we disagree on that . From my point of view, there is no requirement that an edge be permanent.

 

[MarkBrown]

Quote from piezoe:

Well, we disagree on that . From my point of view, there is no requirement that an edge be permanent.

ltc failure was they had book smarts and not street smarts or the perfect combination of these two. they also suffered from pride because they would not admit their formula was flawed. plus they did not prepare for failure or they would have survived. many others prepare for failure before they even think of success, and your practice failure scenarios in advance.