Originally Posted by pa18 Why would an indicator repaint, couldn't the signal be a new one? |
I'll take a stab at an answer here, though I am not the most qualified to do so.
Suppose that an expert advisor, .efs, .eld, etc. were programmed to signal a long trade whenever price retraced a fib% after making a new high. On one bar, price retraces 23.6% of the move, which is the minimum amout required to issue a signal. A green up arrow is printed beneath that price bar by the program to signal a long trade set up. Then, on the next bar, rather than price resuming the rally, it continues to decline, and reaches a 38% retrace of the move. This is within the parameters of the program to signal a long entry. The green arrow that had been printed under the prior bar now moves, i.e. repaints to the subsequent bar. Is this a new signal? Or is it the same signal just from a new level? Does it matter? I don't know how to answer that, as I do not use any such programs. To continue with the example, price continues to fall, and it hits the 62% retracement on the next bar, and the signal again repaints from the prior bar to the present bar. On the next bar, price retraces 86%, the maximum retrace allowed by the program to still issue a buy signal. The program repaints the buy signal on that bar. Price continues to fall, retraces 110% of the move, thus invalidating all the signals that had previously been painted on the chart. Often, there is no sign left on the chart by the program that it had issued what amounts to four losing long trade signals in a row. Sure makes back-testing tricky for the user, but it probably makes it much easier for the unscupulous software seller or systems vendor to cherry pick examples for their sales literature and web sites.
None of this is meant as a criticism of The Rumpled One. From what I have seen of his work, he has some fine indicators for those inclined to use them.
Best Wishes,
Thales