The last time I did 5 bars was the last time I will do 5 bars. wait -- are we talking about the same thing?
%% Well its true- 5 days in a week, i enjoy weekly charts more than 5 minute candles. But if//when you find a top trender like QQQ or IBD leaders, [oil in a good doWntrender...,.....] dont be surprized when it does 6 bars , 7 bars, 9 candles/bars or 7 months/candles + more. NOT a prediction; but when you got in @ 2 or 3 bars//candles. + say you like top trenders,9 months///candles,bars trend happens..... Especially when top trend volume is much better [200 dma.....or 50dma ] than counter trend volume...,,......; of course factor in stock seasonals ,elections+ other factors..................................................... NOT prediction; but i like pay on 6 days//months better than 5. I seldom sell on 5th day/friday; but not a prediction Thanks.[Edit remark; top trend comment, which differs from an average trend]
%% Good ones. Average age in UK=78 years;US= 77, Singapore[low taxes/tariffs]=82 years. Its much less for people that cant respect borders. Its higher for woman, all 3 nations...... Numbers are not exact nor are they predictions; but still helpful.
As a long time trader (swing/momo) of classic chart patterns, my avg time to hit the target is about 4 weeks. I am not 'seeing' into the future, I am projecting the potential move based on the 'measured move' which is typically an equal distance of the price range inside the pattern. Hope that helps answer your question.
Try your TA strategy with tme based exits. Longer may not be better if MAE and/or winning percentage is low. Shorter could be better where time in trade and drawdown are low and win % is high. Giving back a large percentage from max profit is holding too long unless that is part of the strategy. Each instrument nay behave differently as well.
I guess I'd like to push back at the OP phraseology. There is no TA, FA, AI, or XYZ that sees into the future - five seconds, one minute, five minutes, five days, whatever. Calculated risk is all we can hope for.
There are techniques based on linear prediction that project values into the future based on past data. However, I doubt in trading that they would give anything more than 50% accuracy in results, but I haven't studied the technique too extensively.
My $0.02... TA is just a set of tools to understand the story of the chart you're trading on. That's it. In the right hands, I believe it can be a powerful way to help inform trading decisions, BUT... (a) not the only way to inform them, and (b) not a magical crystal ball that can see XX days into the future. I mean, the chart gives you a story that you then have to interpret and being right about what it means, the direction of the next move, is hard enough to do on your own consistently. A better way to think of TA, for me at least, has been in terms of key price levels... because the duration is a factor of that, the volatility of the security, and the timeframe chart you are watching/trading from. That way I can be a bit more objective. I remember one trade in particular that turned out great for me, but took much much longer than i expected. I could sit here and say that my TA was so good that I saw months into the future, but truthfully I didn't know how long it would take. I just thought I was right about the direction and thought I had a good idea of where the next level of resistance was.
I think what you are referring to is linear regression - please correct me if I'm wrong. The OP's thread was referring to the predictive capability of Technical Analysis - which for 99% of the ET Members reading this means the standard TA studies that they get with a free or inexpensive charting platform. Which is kinda worthless. The better thread would be something along the lines of: "anybody out there using predictive modeling on market data?". Then you're getting into the good stuff; neural nets, machine learning, and all sorts of statistical modeling solutions. This is a much more robust approach than a MACD study on a chart. Much. But again, there is no legitimate "predictive" capability. I will say it again; calculated risk is the best you can do. And that's good enough to make bets on markets. The reason that the future cannot be predicted is because studies based on historical data assume steady state conditions in the markets. Looking forward based on history assumes a constant. And as we know, markets are complex systems.
Classic TA (old school) is based on chart patterns & volume as taught by the founders, not indicators. Measured moves are used to identify a potential profit targets when price closes outside well defined classic chart patterns. It does not work with any real certainty so good risk/trade mgmt always takes priority. The breakout from the ES rising wedge to the first leg down is 215.33 points (A). An equal sized move down of 215.33 points is projected from the reaction high on leg A. Down leg B moved another 215.33 points as projected from down leg A. This is what is called a measured move in classic TA, it does not attempt to forecast the time it takes for the moves to play out - although they are typically strong directional moves. I figured a picture is worth a 1,000 words. Only an estimated 3% of all market participants use classic TA - I like that, not much competition.