it is never wise (repeat NEVER) to add to losers as part of a strategy. Show me a SINGLE successful trading strategy that works well by adding to losing positions there is none!!! Now, this does not mean you cannot every now and then add to the position if the intial position went slightly against you and probabilities are convincingly high enough of a rebound (and this I only recommend to someome who has traded and watched a specific market sufficiently long to know it inside out). But suggesting you can build this into a strategy demonstrates that you may know about building code but surely dont know how to trade. I can only urge every beginning trader to be bold to add to winning positions and rather cut losers at the initially set threshold, no debating or afterthoughts.
However, this remains only a theory unless it it put under rigorous testing. My experience is that people who add on even partially, wind up pounding several system trade statistics they may not think about. For example, everytime you increase your positions/leverage, your average drawdown per position will jump (this is strongly tested, not guessed). No matter how it is sliced, adding positions is the same as increasing your leverage. And all the corresponding positive and negatives that come with it are there, whether one is trying to do your stated objectives or not. As someone said, there are no magic beans. Opinion or belief does not make something any more value than other opinion. Statistical testing is what tells someone that their method is rigorous and well designed. If you want to add truth or value to what you posted. Then say "here is a random sampling of 300-400 trades where this practice ocurred. And analyzing the results indicate that..."
Fully agree with your primary point in that opinions vary. What one person consider reliable or good returns another may disagree. It depends on personal financial objectives. There's 2 factors that affect your opinion and make mine different. If you test this on a single strategy, it's less desirable. But if you do the same on non-correlating strategies where they balance each other, you get much prettier equity curves. And for your test to affect anyone elses opionions (if that's your aim) you need to say if you did any of the other steps like throttle adding to positions or filtering out some of the entries. As just one example, I discovered early on that all my biggest losing trades (due to my limit) happened in exact correspondence will all the big news reports (this was on Forex). So every time the market reacts to major news event, the market oftens move so far, so quickly that adding to losers is very dumb. However, there's 2 ways to handle that depending on your taste. 1 is figure out a way to filter out those events while keeping in mind that some major news events are "unplanned" like terrorists hitting the twin towers. If you're added to a losing position in a catastrophic situation like that then kiss your assets good bye. I personally found an automated way to filter all those major events out which I refuse to divulge to anybody, sorry. The other way to handle that is to measure the velocity of price change and avoid entering or adding to positions during a fast moving market, still I find it works best to recover by adding all the skipped positions when it pauses to take a breath. None of this is fool proof of course. It's a game of playing the law of averages. However, some traders just cannot sleep at night unless we have 90%+ of our monthly or weekly results come out as winners. Others who can live through a majority of weeks or months as loses will do well with trend following type strategies. But it seems better for everyone to mix trend strategies with other strategies. Ohh la la. I think the new fronteir is combining non-correlating strategies on the same markets and re-balancing the asset allocations based on performance or other factors. Successful traders do very well with fixed percentage asset allocation to strategies, for example. There's many, many other ins and outs to this. I don't think any anyone with less than around 10 years of study, research and trading experience could even attempt 90% + winning strategies and smooth equity curves through portfolio trading. But opinions are like noses, as they say, everybody's got one. Sincerely, Wayne
By the way, my personal observation about myself and many traders is that we initially come to trading with an almost gambler mentality. That is we're looking for a huge windfall payoff. With that mentality any measely strategy that only makes 5% to 10% per month never grabs any interest. So like travis mentioned, I was originally hoping for 3 to 5 digit returns per month. Ha! It took years of testing strategies and actual trading to finally realize that wasn't realistic. The big institutions and others profit not from big windfalls at intervals but many small consistent profits over time. That's true whether it's market making, arbitrage, you name it. Once traders make that realization and look for many small profits and consistent returns, things turn around because that's far easier to find than any get-rich-quick strategies. Of course, it's still far from easy and requires learning money management, risk management, portfolio balancing and requires tools that can do all of that. Sincerely, Wayne
"...almost gambler mentality": it applies to me. "...measely strategy...": correct, because we don't realize we can build 20 of strategies and use them simultaneously. "...It took years of testing strategies...": same for me - I realized, after 5 years of testing, that it's better to have 20 strategies that promise less and deliver it all, rather than one that seems to rarely fail, while it's just over-optimized. "...money management, risk management...": right, it's not like because you have a system that works now, that you can just invest your whole capital on every trade it makes (something I did for all of 2008).
I disagree with asiaprop on only one point: that it's never a good idea to add to losers. There are scenarios where it makes sense. Chan was just talking about this on his blog, actually... his principal of latest information. http://epchan.blogspot.com/ If you're trading a mean reversion strategy, then you might find yourself adding to a losing position. But that's still night and day away from any strategy that adds to a loss "just because" you lost. That's a martingale, period.
FYI: Martingale is a simplistic form of a mean reversion strategy per comparing these wikipedia definitions: http://en.wikipedia.org/wiki/Martingale_(betting_system) http://en.wikipedia.org/wiki/Mean_reversion_(finance) As you know, mean reversion means you're expecting the results to eventually somewhere near to the average. That's all martingale does. My belief about what makes martingale truly a martingale rather than just a mean reversion system is that you always add to losers without limit. That's why you run the risk of wiping out your account. In any mean reversion system, it makes total sense that the further price gets from the average the more likely it will snap back and more reason to add to a position. In fact, my research shows that prices revert to at least 50% of any advance or decline always. The main problem with martingale is that sometimes in a big move it can take weeks or months to get back to 50% so if you add blindly, those big moves can wipe you out. I personally use adding to losers it ike you say for mean reversion but other ways also by using loss limits and filters based on trend, countertrend, chop, channel, etc. That's because in an uptrend the prices tend to avoid returning to the mean until near the end of the trend--bad time for simply adding to losers. However, during trends prices consistently return to 60% or 70% of each advance. Since my strategy measures each advance and retracement, then, in reality it's looking at a form of support and resistance in the mean reversion. When the market is in channel mode, it usually retraces 80% or more of each advance or decline. So you play adding to losers differently. In fact, you could almost measure what market "condition" you're in by how bar prices retrace %-wise before reaching new highs and lows. My experience generally says that a trend, for example, is totally over when it fully retraces 50% from the beginning and that often leads to a "chop" market condition for a while. You're right. Adding to losers as a strategy all by itself, is dumb no matter how you slice it or what you call it. But in combination with mean reversion, loss limits, support/resistance, retracements, money management, risk managements, asset allocation, multi strategy per symbol, (the list goes on)... Then it's just another element in a big puzzle. Wayne
so true. It takes a while to realize it is all about Risk-Adjusted Reward. You can make 100% a month on leveraged instruments. All you have to do is be able to deal with 100% drawdowns several times a month
This is not true. I have watched MANY systems of others crash precisely when a large move happens, precisely on this assumption. Unleveraged stock trading, maybe. Leveraged trading, your assumption is only true until it blows up. The market does NOT revert 50%. Leveraged trading cannot afford to wait 9 months. The normal market is fine, but every year or two, it blows your strategy out of the water (leveraged trading).
fyi, I did not find much value in Chan's book and I know for a fact he has never made a single cent trading, bouncing from one job to another. Just so you know on what source you rely...I can recommend Perry Kaufman if you are looking for some sort of strategy setup. You wont find any strategies that work in there (where do you, its your own work to do) but its a solid writeup of strategy building and TA.