Good thread, interesting other than the fact no one has a clue on how you come about your predictions so there's no value here for anyone. So far the results are no better than a coin toss whereby on your thread introduction mentioned better than 60%. Do you have any comments on this? Was the journal's purpose to prove something?
Not being critical, just wanting to see where there is value via feedback. It is possible your method can work more efficiently, it's just that maybe atm it's not working. The reason being, atm I'm also struggling to make sense of mkt direction, we seem to be in a chop type phase where it whipsaws around and lacks clear direction. Although on the surface there is no clearcut visible whipsaw, what I'm experiencing is selloffs toward the end of each day. The mkt may look bullish but deep down there is conflict imo.
it is Machine Learning model (I won't dilute the details here). Unfortunately the model was put in 'live test' during one of the worst volatile time frame (early of Oct), and the model is not able to make a good prediction as those are not "normal" price movement from historical perspective. This is always the case in real life trading, you always get into the 'wrong' time after you develop the model. I expect the model prediction will get better when we look into in longer time frame.
Very good. Mkt signals atm are indeed confusing, I've stepped aside, shut down my trading, preserving capital. I'm not even short.
Just noticed 46/91 Shouldn't 91 be divisible by 4? I'll let you check back and see where you started with odd numbers.