I have been working with multiple screen trading strategies like Elder's Triple Screen and some other derived techniques, and they have proven successful at least 50% of times, which is just enough to yield profit when used with strict money management rules. For a while i have been trying to simplify everything and came up to some very very basic single screen strategies, that are also profitable 50% of times (ok maybe 40% but still enough). I feel that sometimes, when you add another timeframe is like adding another indicator, it clutters up information. On the other hand, when trading with the trend, having a higher timeframe trend keeps you out of trouble, but it also avoids trades that would have been profitable .... How many of you use single screen strategies and why and how many use multiple screeners and why?
Your question is too broad and is framed backwards from my point of view. Would you not be better to develop a strategy that included a primary time frame and then decide whether multiples were a help or a hindrance.
You don't understand. As i said in my original post, i am already using a multi-timeframe strategy (derived from the Dow Theory) but i am also using singles. I am interested whether people use singles timeframes or multi-timeframes on a statistical basis. I don't really need advices on this post, thank you...
Wavelet decomposition offers the possibility of optimal denoising and is based on multi-time-frame analysis. Unfortunately it's after-the-fact denoising so I'm not convinced it's useful for an actual trading strategy, but some people still seem to look upon wavelets as a potential trading tool.
I use a higher timeframe to help judge the strength of a signal, which for me translates into position size, and profit expectations for the trade. This is by no means scientific, but it's a guide that I've found helpful as part of my intuition when going into a trade.
It depends on what you're trying to do. But surely the information utilized as part of a strategy based upon bang up to date information must tie to some degree with information from the past. Therefore, denoising is useful assuming there is useful information out there to be found. The more tricky question then to ask is how does one define that noise? Thx D
Single timeframe works fine for me. Isn't all about price anyway? Does the market know or care if price moved a point on a tick or a 15 minute chart? On one chart, a bar could be red and on another its green. All the information of the market is in all charts regardless of timeframe.
Taking trades off a 5 min chart and looking at a 15 and 60 mins to obtain ''confirmation'' is imo utterly useless. what i personally do is I look at the Daily ATR and ADX, and this allows me to figure out which systems ill be trading on a particular day. It is pretty dangerous to trade a trend following system when the market is not in a trending phase right? so if daily ATR or ADX is low, ill be trading a fady fade fade reversal system.
found at wilmott.com, posted here without commentary: "I think wavelets are overrated. Wavelets provide optimal representation (in the nonlinear approximation sense) of piecewise smooth signals. If your data is piecewise smooth (like e.g. images) then it would be hard to beat wavelets. If not, then there is NO reason to believe wavelets will be useful at all."