Yea, I've always wondered about this. If you base your bias on the NL and it's deeply negative or positve, you'll definitely miss out on the rebound because it takes so long for the NL to get back to 0 and confirm on the other side. I don't have a way to deal with this. Either we miss out on the rebound and only trade in the direction of the NL on a longer term basis or we start peeling away, like mav said, and start to analyze different layers and look for early confirmations on a firming NL rather than an already confirmed NL. I feel the latter is what Mav and everyone here is alluding to...but that comes with a lot more ambiguity as well.
The ACD Method
- Thread starter sbrowne126
- Start date
Yea, I've always wondered about this. If you base your bias on the NL and it's deeply negative or positve, you'll definitely miss out on the rebound because it takes so long for the NL to get back to 0 and confirm on the other side. I don't have a way to deal with this. Either we miss out on the rebound and only trade in the direction of the NL on a longer term basis or we start peeling away, like mav said, and start to analyze different layers and look for early confirmations on a firming NL rather than an already confirmed NL. I feel the latter is what Mav and everyone here is alluding to...but that comes with a lot more ambiguity as well.
Use derivatives. I have 2nd order and 3rd order derivatives for NLs. They are VERY fast.
I was just going to post that, but had to exit the gmail app and load in Safari before I could write at a decent pace. Shan talked often of 2nd order derivatives, I clued into 3rd order derivatives after a recent post of yours. Was easy to see your point since I have studied options.
Hint: it's the rate of change, or the rate at which the rate of change is changing that matters. That's where you get the early clue.
Its funny you post this, because I have been thinking about this for sometime.. A guy I know uses a proprietary variant of the mcclellan oscillator. Now many people throw out the mcclellan as being only effective at certain times. The guy who i know uses 2nd and 3rd order derivatives of the oscillator to great success. For example he measures the slope steepness of the oscillator and this tells him the strength of trend along with what would also be needed to reverse that trend.
What this means is that you can have an indicator that 'looks' extremely strong but, its only by observing the derivatives that you can confirm this. I am quite certain something similar to this would work really well with the number lines.
Now you've triggered my creative juices - another sleepless night 
Honestly, I don't use any data in raw form. ALL my data is transformed.
In principle, from the minute we extract the daily score from nominal price we are already working with transformed data. I look at the daily scoring method as a way of 'classifying' the various phases that occur in all markets.
Yes, this is true. You are right. I like to keep peeling away at data until I extract the most amount of noise and get to the real information.
The other byproduct of daily scoring is that the data is de-trended and 'stationary'. Trying to make OHLC data 'stationary' can be a nuance as it is highly influenced by lookback period used etc.
Daily scoring alone is a massively powerful tool and makes the all other methods of OHLC analysis seem redundant. Since I using ACD I have never looked at another price chart the same since.
It becomes even more useful for 'price action' analysis as you have mentioned numerous times already on this thread. Especially when compared to the other methods of monitoring price action such as highs vs highs etc.
Every time I read a thread on other trading forums where people have been discrediting ACD since it doesn't stand up to a backtest, I just have to laugh. These people really aren't seeing the forest for the trees...