I guess I can add a third situation where a stop can be helpful -As a psychological trick, sometimes having a mechanical rule to get out helps one feel that 'its the system fault' or 'see how disciplined I am, taking this loss and following my rule' and decrease the amount of psychological damage the loss can generate
I'm afraid most people using technical stops don't have an edge on them simply because they have become too widely promoted(highs, lows, round numbers, common TA levels etc) so they are just paying more in commissions/bid ask spread when they could be better off working more on their analyticall skills and cutting their drawdown fears through position sizing
I've found that certain trend following methods produce slightly more profit with an optimized stop....... on past data. However, tight stops ruin it in a big way. As you say, position sizing is the real deal. If the daily range will blow out your account, that would be a hard life. I use stops more as a safeguard for my own psychological weakness. It's a way of making a decision in advance, rather than on impulse.
Agree with most everything you say -- the process I described has a definite and undeniable skill component (accurate assumptions etc). But then, this is also why it works -- any time one expects to produce alpha via methodological process, one must also have a credible explanation as to why such alpha is sustainable (not readily competed away). If any Wharton school grad could crunch the numbers and replicate the strategy, there would be no edge worth applying. Re, holy grail, agree with that too -- if I found an investment with that level of return and commensurate conviction, I would be happy to put a big chunk of net worth into it. The devil is in the conviction details.
I'm not sure the value of stops can be discussed universally, i.e. outside a specific methodological context. With some methodologies they work well; with others they don't work at all. Sort of like discussing appropriate debt to equity ratios for a business; there is no one size fits all, it depends on the business. I would say this is not "an argument against stops" so much as an argument for why stops do not work for you. Consider this counter-example: A mechanical trader creates a profitable trading methodology based on intermediate term momentum swings. In both backtesting and post-analysis of live trades, this mechanical trader discovers that, on the whole, if a position moves against him by X volatility factor or does not become profitable within X days, the odds of a follow-through signal have degraded enough that he is better off simply exiting the trade. What this trader has done is built a profitable methodology around exploiting risk:reward scenarios on a mechanical basis. Part of what makes the system profitable is cutting down losses by immediately exiting when a signal has become degraded -- in other words, the edge is in the backtested methodology and cutting losses through stops is a part of this system. I am not a mechanical trader; the idea is that the value-add of stops to a methodology is wholly dependent on the context of the methodology itself. This difference is not a matter of degree vs a fundamental conviction approach; it is a different way of conceptualizing the problem entirely. Not better or worse, just different. Which goes back to the original point: It is hard to discuss the worth of stops outside methodological context. This is true -- a poor approach will find a way to bleed out over time, whether fast or slow. But this is not a knock against stops per se, just a recognition that good money management cannot mitigate bad trading. I don't scorn Rogers either -- he has been too successful, and what he does seems to work for him. I do cast a wary eye on "unshakeable conviction" types though -- their style could not work for me, much as my style could not work for them. I believe too much in fallibilism; I see the world as too complex a place; I am too in touch with the question, "What if I could be wrong?" Three examples of the Rogers mold gone bad: Bill Miller, Bruce Berkowitz and John Paulson. After a string of excellent years, those guys took their unshakeable convictions and rode them right into the dirt. Not saying all conviction traders do this -- just that it's a risk I find personally unpalatable. Right -- saying "the stop itself is an edge" is another way of saying "the stop is an intrinsic part of a profitable methodology." Re, Thai baht to corn etc., the easiest way to normalize risk points is as a function of price action and volatility, i.e. chart levels and average trading range. We do this as a matter of habit with position sizing too; for us a "1.0" position size means planned risk of 50 basis points, appropriately adjusted to the volatility of whatever vehicle is being traded, which can result in very different notional exposures of one vehicle vs. another (three natural gas futures contracts might work out to the same planned risk as thirty corn futures contracts etc.).
Perhaps, but such approaches always struck me as childish. There is so much written on the need to be disciplined, to follow one's rules and have confidence in one's approach etc... I would say this is kid's stuff, akin to saying "if you want to be a professional athlete you have train regularly and not eat donuts."
If someone is losing money in the market with the use of stops, I'd say they are likely to lose even more money without them (frying pan into fire). Also, the stop itself is not the edge, as a risk point says nothing about entry point selection, vehicle selection, position size, portfolio correlation, equity curve management etc.
A bad stop is where you exit just because the price has moved against you. A good stop is where your exit because the trade expectation has moved against you.
On long-term vs short/medium-term: 1. Risk is identical, because you adjust your size and therefore your risk based on your risk/drawdown tolerance, not the upside. Thus, what differs is the returns per unit risk, not the risk itself. 2. Return per unit risk is a function of how good your trades are. By definition, the best trade in a year or decade is better than the 2nd best trade, and the top 5 trades over 10 years are going to be much, much better than trades 6 to 500. So an approach which goes for the top 5 trades, rather than the top 500, will perform far better, other things being equal. 3. OFC, other things aren't equal. The longer the holding period, the more backing & filling and corrections you have to endure, which reduces CAGR. That's one downside of longer-term trades. 4. On the other hand, short-term timeframes require much more timing ability. In theory, the perfect trader who can buy every low and sell/short every high, will do better the shorter-term the timeframe he trades in, simply because there is more price movement to exploit. 5. Major benefit for the ST trader - the more trades he makes, the more times he gets to apply his edge, which means more profit and lower DDs. But, in the real world, this is offset somewhat by the lower average quality of trades taken. Hindsight bias applies to all timeframes, it can't be used to dismiss one type of trading and not another. So, overall IMO it just boils down too much to specific real-world factors, it's not an argument that can be settled by theory. Buffett before he ran into size constraints, said he could make 50% a year on small/medium size, but he had occasional big DDs e.g. 50% in 1973-74. Soros in his first decade did 40% after fees i.e. 50%, and had lower DDs. Renaissance made even more and had lower DDs. So overall it seems like the shorter-term has superior performance. However, Buffett had to pay taxes, Soros/Simons didn't. Most traders have to pay taxes. I would however argue that Buffett is not the pinnacle of what can be achieved in long-term speculation - Buffett only operated in US stock markets early in his career, whereas clearly the biggest booms and busts were outside the USA, and adding commodities, currencies, emerging/frontier markets etc would have radically increased the opportunity set. Buffett was more a value investor and was pretty risk averse, he didn't like to speculate. Soros like Druckenmiller was not really a long-term speculator, Quantum only did that while Rogers was there - once he left their risk-adjusted alpha dropped radically (50% per annum for a decade under Soros/Rogers against a flat S&P, with moderate DDs; versus 40ish under Soros alone, and Soros/Druck, with much bigger DDs). And I think the Renaissance edge was somewhat structural and is vulnerable to increasing competition and size constraints, whereas macro and stock investing isn't really hindered as much, since the opportunity is provided by errors in human psychology (i.e. permanent inefficiencies) rather than the competition not yet having figured out your edge (temporary inefficiencies).
Not quite that cut and dry... The ability to successfully trade shorter time frames allows for expansion into longer time frames with less risk. An analogy from poker: When the stacks are very deep -- say 200 big blinds or more -- it becomes possible to put 1 to 2% of initial capital at risk preflop... "pyramid" exposure on the flop and turn... and win an opponent's entire 200 bb stack by the river. This is an escalation process where the original risk was minimal, and exposure increased only as prospects for gain dramatically increased. What began as a carefully timed trade expanded into a large investment -- but only under confirmed favorable conditions. This is a very different way of going about things than, say, putting 25% of one's stack at risk from the start. The ability to trade short-term does not preclude long-term profits; if anything, it enhances the opportunity to generate them (if one is sufficiently versatile). Also, re, "going for the top 5 trades" -- Ken Grant, risk manager to some of the best managers in the world (Jones, Bacon, Cohen etc), relates in his book "Trading Risk" that many top traders have a 90/10 profit ratio, meaning 90% of their profits come from 10% of their trades. Does this mean that all trades beyond the 10% are superfluous? No, because the other trades are a part of the process: Testing the waters, jockeying for position, easing into scenarios as they unfold etc.