Now for the month of February we still have several more sessions to go even though the month ends tomorrow we need to wait more than 1 week to finish evaluating how did it go for it (so we can see if the predictions this week work next week or not). However running things so far it looks terrible for the NDLA (the market corrected 10% so it is no surprise) also because of the incredible expensive implied volatility we didn't officially play all the signals because they were outside the recommended trade parameters (i.e cost <= $1.5) However even if we only played a few let's look at the performance so far:
Full Precision: We got 5 signal in total so far and only 1 of those was correct, so precision is 20% so far (more than terrible).
Recall: We got a total of 6 moves of 1% or more so far, so recall is 1/6 or 16.7% (very bad as well)
Partial Precision: Incredibly we got 3 partial successes (moves of 1% within the 5 session period) so a 60% value for this (much better than full precision)
As you can see the performance so far has been terrible so to add salt to the injury let's compare with the Naive strategy even though, again because of the cost, we should not have done it at all:
Full Naive Precision: 6/13 or 46.15% (very low but actually higher than the expected historical 30% threshold).
Naive Recall: 100%
Partial Naive Precision: 9/13 or 69.2%
So as you can see in times of high realized volatility the naive outperforms the NDLA (at least so far) however the point is almost moot because the pricing of the 1% up binary has not been under the edge threshold for a while now ($1.5 or less)
There is another discussion worth having here and that is if the under-performance of the NDLA continues will that signal the end of the bull market cycle ? The reason I word it like that is because the NDLA is looking for patterns that have been very common during this 9 year old bull cycle, so if those patterns start to fail it should be sending some sort of signal to us that the bull market might be ending. That being said we have only have 13 sessions in Feb and the combined performance for the year is still very good but I would really like to see these patterns working well consistently again very soon. The good thing is that at least the last 4 signals generated (which are not included in the analysis) seem to be already fulfilled (at least partially).