Machine Learning
Last updated
Last updated
In the recent 1.0.7 update, we released a machine learning feature to the baseline.
The baseline will track other popular indicator values in real-time (RSI, CCI, ADX, and some other popular indicators). It will scan all of those indicators and all of their values for the entire past chart history (looking at all of those data points from years of past data), and then find what price did at those specific values. Did it go up, go down, or move sideways?
The indicator will note that and essentially create a "map" of what price did at what indicator values. Which is this screenshot right here:
Once it creates that map, it will look at the current bar of where price currently is. We take the current values of those other popular indicators and look for the closest data points from past history compared to the present. So it will basically try to find data points from the past that look the closest to what it is currently.
It will then compare that to what price did back in past chart history with multiple instances. It will then compute the distance from every previous point, and then average out all those changes.
This essentially can give a great prediction of what price is likely to do based on the previous data values and observed trends.
The cool thing about this though is the more that this indicator runs and the more data it receives, the more accurate it's going to get in the future. Which is pretty cool in itself.
We completely redid the baseline for this new update. Essentially it works like this. The darker the green, the more the machine learning thinks the price will go up.
If its a lighter green, it still thinks it's going to up, but the probability is just a little less.
Same thing with the red. The darker the red, the more sure the machine learning thinks it's going to go down.
Then grey is basically consolidation.
The trend table has a new "machine learning" metric. The value will update in real-time and will have a corresponding color depending on if it's bullish, bearish, or consolidating.
Dark Green => 75% and up
Green => 50% and up
Light Green => 25% and up
LightRed => -25% and below
Red => -50% and below
Dark Red => -75% and below
The new baseline can be used in a lot of different ways. I think it will be best used for confluences. For example, if there is a buy signal from the indicator, make sure the baseline agrees with the signal.
You can even go on larger time frames and check what color the baseline is on a larger timeframe. Once you know the color, you can then go on a lower timeframe and only follow signals correlating to that color.
Another way it can be used is for take profits. So for example, in the image below, the indicator gave us two long entries. The first one worked out great and we got both take profits. The second buy signal only gave us one take profit. After the first take profit was hit, the chart tanked and we ended up exiting for a loss.
If you look closely, the new machine learning baseline turned red right after the first take profit. which could be a great indication that the chart is about to fall and we should just take our profits there.
There are a lot of different fantastic ways we can use this indicator. I'm excited to see what we can find. The machine learning feature is only going to get better over time. Very exciting stuff.
