All the responses given above by others are good advice. Here are a few more things:
- A trend line or channel you draw is not necessarily an absolute thing. You should take it as hypothetical until you see price conforming to it. That means you may also have to fiddle with it until it fits. You may have to just discard it and try a different one. You'll need to abandon what isn't working.
- Trendlines are like any other support or resistance line -- many, or most, will fail, but when price turns it is usually at some support/resistance level, so it pays to have looked at many possibilities. Trendlines are support/resistance for pullbacks and rallies within trends, so think of them the same way.
- Trying to draw a parallel channel will often tell you if you're right. A good channel will contain the tops as well as the bottoms. You can adjust both lines until they are capturing price well on both sides.
- Sometimes you can get the right line by drawing the other side of a parallel channel first. So if I am trying to find the rising support line across the bottoms of bars, sometimes I will see if there is a good rising line across the tops first, and then make a parallel channel that touches the bottoms. This can also sometimes give you a valid lower line before you have enough lower points to connect up (for example, connect 2 points on the top of the channel, draw your parallel from only 1 point below).
- That said, not all trend lines have to be parallel to something else. "Wedge" and "triangle" formations have lines that converge. You can also draw lines that don't have another one paired to them.
- Look to a short-term moving average to guide you in the slope of the line. You can put a 20 EMA or something like it on the chart and draw lines that run parallel to it.
- Be willing to try weird lines that you aren't sure will work. If they don't, who cares? Al Brooks wrote somewhere that you can and should draw lines connecting any 2 points and see if they work out.
This is a lot of stuff and it may not help too much by itself. Sorry. The best idea is to get into it and just draw some lines and see if they work out.
Sometimes people will criticize using trend lines because they seem arbitrary, or you have to try many before you find one that fits. Or, you may not find one that fits until the move is over. These are valid points, but with practice you can find lines that do work -- ahead of time.
Last thing: there are some threads here on futures.io (formerly BMT) where people are posting their charts with good trendlines and channels. You can see what they're doing to help get the hang of it. You can also look for resources by people like Al Brooks or Mack on the internet. Here's a good futures.io (formerly BMT) thread, and there are others: