So i've taken a look at the paper, they mention that their best result was in a highly volatile market which obviously the cryptomarket is... i'm curious too now how well this would perform in this market. In theory i guess you could even change the trading days into trading minutes but for that I think the algorithm would need to have to take different things into account like support, resistance, rsi, ... . I'll first see if i can run some tests with old coinmarketcap data when I find some time one of these days and i'll report back.
You are viewing a single comment's thread from:
I am glad that you are now more interested to it ... feel free to contact me any time here or at my facebook account ..
The professor created all the algorithms including PAMR in OLPS ... you can find it here : https://github.com/OLPS/OLPS
Your help is needed and highly appreciated.
Keep going champ
any news ????
So I tried the algorithm with the following tickers: "ETH", "BTC", "DASH", "XMR", "STRAT", "DGB", "DOGE", "STEEM", "XRP" (so one portfolio containing all of these). It seems that the algo performs worse than a uniform constant rebalanced portfolio. Check the image below. So as ashr mentioned this does not perform that well ...
http://imgur.com/A7YHKWH
Thank you for your time and effort. at least you tried....
Thank you so much
Sorry was busy last days, I'm looking at it right now!