Trading and Medicine analogy (complexity, uncertainity, technology) - part 1

 Avi - 

My Cryptocurrency trading setup.. When I used a  profitable + generalizable strategy (eg. RSI overbought and oversold cutoff) and let it run free 24*7 on fully automated basis, it gave some profits but when I use same strategy and add some personalization / precision (same RSI based but sell only if it's in profit, and also when noticed bitcoin dominance then choosing high volatile tokens where percent change is more on both sides up/down) then profit increases drastically and also losses reduces drastically too. Are we missing the drastic/tremendous benefits of drugs because of not having the right information and analytics on top of it?  Or we are just loosing a lil.? 

May be we are loosing in ratio of NNT, as if More precise and beautiful action of a drug, better optimization and  benefit scope and may be we are doing fine without optimization as we haven't seen how much gap is there in the approaches.. 

Probably looking at evidence into cancer care will give info on variations in impact of drugs by following personalization strategies as they are most advance in that afaik.

( I had just begun algo-trading using ActionScript and google sheets that i made over a weekend)

these values and terms I made myself with some analytics, not sure if they are defined and exist already, must be I guess.

RB - Good question and hypothesis.  The big question is how do we test it.

Avi - It's recommended to develop own strategy and keep modifying and improving it. No strategy can perform best/optimal in all markets (different time in a year or even a week sometimes).. I created my own based on freely available knowledge base that i could study and some hypothetical experiments tested later with practical experiments in real market. 

Here in the excel sheet I have programmed to fetch real-time data and developed my own indicator named co-efficient, it is probably similar or in relationship to B (beta).. Now having a quick look at this sheet helps me make better choices for faster profit. It doesn't help choose profit/loss, but whatever happens will happen at a faster pace in the chosen token based on co-efficient values.. A programmed script also analyzes the real time data as I would do manually and sends me alerts as mobile msgs so I don't need to keep looking at data, rather just check when there is a buy alert and decide to go ahead or not.. 

In a human the algorithms may analyzes objective and subjective data generated by human body (sensors) and a doctor may only give permission to execute a best possible chosen Rx/Tx..  But there are a scope for a few errors and combined together they may create havoc.. 

1) population level data - algorithm ( all data in a range +/- and we go right way or wrong way makes us get small/big error) 

2) human level - algorithm input (sensors - natural and artificial) 

3) decision maker - algorithm+human analytics based output/decision (doctor's choice/decision)

Avi - So now how do we test it/ minimize errors? 

1) by making corrections before it's too late. By Making smaller corrections to keep the course of therapy more precise to the desired outcome without loosing too much against uncertainty

2) developing sensors and their analytics to make the first point a reality. 

The point 2 reminds of a line @Dr. RB sir repeated a few times during electives - we need micro level sensors data and macro level data and merge together to decide best possible way ahead.

RB - Is this what every individual doctor currently does in his own self styled intuitive manner?

Yes sir.. But intuitive is not a more of data focused approach neither the outcome tested and corrected in real-time.. 

(For a system realtime may be also every Nano-second, and may be even hours, we just need to zoom-in to the right ratio).. 

(Idk if the real-time term use is technically correct here).

(Intution is data based - much subjectively, less objectively)

So corrections are not non-existent but not as high precision as we are thinking as possibility by the hifi system.. 

Q - do we need such high end precision? Given the facts that 1/3rd medical error deaths, 85% dx healed even without any intervention, most tx are emperical, etc.

RB - We could reduce overtesting and overtreatment in those 85%?

Avi - Answer - Yes need precision

Not just on optimizing tx but also finding better tx and finding better dx and most importantly ensuring their access because if this exist only in ivory towers then non-existent is the real reality of world.

Yes sir.. That can be reduced with only all the data we have now openly freely accessible, just need to be used without even any need for anymore fancy tool other than any Wikipedia or blogger like cdss or tech enabled cdss or at least a conversational cdss

RB - Intuition could be data in quantum reality?

Avi - Anything having the term quantum is very difficult for me to understand..  This term have so many meanings in so many different contexts and so many explanation on web which are unrelated to each other..

RB - Yes it's a theory of everything