Support and resistance (as how it is typically defined) only comes into play in the larger timeframes. One would want to know before entering a trade that there is no significant obstacle in the way to counter the momentum of the market. On the other side of the trade, one would also want to identify a logical place to exit as the action gravitates towards the target area before momentum dries up (or take some money out if it looks like a minor one only and could be taken out, say off the back of a major news announcement for example).
Having said that, support and resistance in the larger timeframes is also a good opportunity to anticipate/predict a trade off too. But this is no different to trading off a break against the longer term trend or congestion breakout back in line with the trend. But I view this as a setup only and any break / reaction still needs to be confirmed and entry timed within the smaller timeframes. How many times do we see candle wicks poking out of a range that amount to nothing more than the market trying to make a move, hestitating and then falling back into line.
As for candlestick patterns, I don't get too caught up with the detail other than to be able to cleanly draw diagonals and horizontals which either touch the wicks or bodies of multiple candles to "validate" the existence of a trendline, support or resistance. Whether one uses fibs, moving averages, round number supports, bollingers, etc. if one can't draw neat lines, then it probably doesn't exist, exists only in one's imagination, or the market is too erratic to enter anyway. I also like to point out that if neat lines cannot be drawn up, it's not to say that nothing is there. It is just saying that the market (money flow) is not focused on that timeframe right now and by zooming out, drawing of neat lines are likely to be found on a larger timeframe. Now whether that timeframe is suitable to one's trading methodology, psychological profile, etc. that is another story. But in summary, I believe it is very risky to enter a trade based on assumptions that one is not entirely confident with or that cannot be validated to some degree.
