This warrants a response. Let me try to provide more color so that my comment -- the way you define 'trend' (or 'momentum') could, in by-itself', be an 'edge' -- is a little more clear.
In trend based strategies, there are two ways of looking at the problem: (a) Have a 'unique' definition of 'trend' and have a simple trading strategy around it; or, (b) Have a 'generalized' definition of 'trend' and have a 'unique' trading strategy around it.
Unique Definition of Trend
Lets take Fama/French 3-factor model as an example. By defining rate of return in terms of market cap and book-to-market ratio they provided a 'definition' for momentum. This definition is unique. Had they not revealed this to the world, this definition of momentum would have been their 'edge'. Like all other things in financial markets, when people start copying a strategy, the 'edge' disappears. So, one has to find a different 'definition' for momentum that would provided a sustainable 'edge'. This is where, in my opinion, all the work is.
Generalized Definition of Trend
A similar explanation can also be given for disappearance of 'edge' when 'momentum' (or trend) is defined based on price in terms of 'higher-highs' and 'lower-lows'. Hence, there needs to be some 'additional' input. I am deliberately being vague here. This 'input' could be part of trend definition, which would make the definition of trend 'unique'; or, the 'input' could be part of the trading strategy, which would make the trading strategy unique thereby providing an 'edge' -- statistical edge.
Because of the two ways of looking at the problem, providing an overarching comprehensive definition of trend is quite impossible. Herein lies the problem for a meaningful debate based on semantics.
Unlike momentum strategies, in a mean-reverting strategy where one is 'playing' co-integrated pairs, the 'idea' itself is not important 'cos these strategies are not impacted by liquidity and positioning issues -- in other words, liquidity and positioning are very important issues to contend with in momentum strategies, where as, they are non-issues for mean-reverting strategies. Because of this, ideas of mean-reverting strategies, even when exposed, continue to work. When people understand this, they will know why 'trend-followers' don't reveal their strategies to others to test!
I would love to continue to discuss this issue should you be interested, but lets get out of this thread!
As a side note (not that it concerns you): there are some posts here which confuse 'mean-reversion' and 'counter-trend' trading. I hope they understand that for one to trade 'counter-trend', one has to first define 'trend'. It is just that those 'counter-trend' traders are betting of the "wind to die down after the kite is in motion" where as the 'trend' traders are betting that the "wind will continue after the kite is in motion". Debating which one is better is just silly!
Regards,
Monoid.