Pulling a channel from end of wave 1 to the end of wave 3 seems to fit very well.
Rejection off the median + multiple rejections during wave 3 adds confluence to the validity of the channel.
Only thing I'm concerned about is how wave 2 and 4 may both be WXY and alternation is not present. however, wave 4 is definitely more complex than wave 2
possible expanded flat to take up liquidity before the move up