Interesting analysis and I think I've reached similiar conclusions when talking to my friends and associates and seeing their results over the last ~5y; high risk leads to lot of capital loss often for them.
I usually have been playing low-risk investments and want to change that, any articles or advice to make independent high risk bets that actually end up lowering overall risk?
I accidentally wrote this comment backwards and I'm on a device that doesn't let me edit very well. You'll have to put up with some nonsense drivel until you get to the part you actually asked about.
----
I don't have much advice on that, specifically, other than
1. Be very careful. When shit hits the fan and you need that independence most, all correlations go to 1.
Related to the above, understand the cut-throat nature of the business. Actually, reading The Poker Face of Wall Street by Aaron Brown might help with this. The difficult part is not winning big; the difficult part comes after that, when you need to convert people's promises into money and successfully walk out with it.
Nobody is looking out for you. It's nobody's job to make sure you get what you are supposed to. I bring this up because I see people spend lots of time concocting advanced technical strategies but then forget the people skills required to get away with it.
2. Learn and play around with as much statistics as you can find the time to. I like reading actuarial textbooks because many of them work in insurance which means they do specifically that.
3. I'm a big fan of Aaron Brown for putting into simple words many of the most important concepts. Read Risk Management for Dummies (honestly!)
In particular, once you start making specific bets, you don't want to have stop losses that prevent you from losing too much money. That's a sure way to lose money. Instead, think like a scientist. Do hypothesis-based trading. "I think that X, because Y. If I see signal Z from the market, one of my assumptions are invalidated and I will reject my hypothesis and exit the trade."
4. Set aside a small amount of money to practise with. Try to invest in ways that expose you to specific spreads, like big vs small market cap narrowing, or service vs product sectors widening.
Try to make these actual bets you believe in. The point of this exercise is that you will, at a low cost, see how often you're wrong and go back to a plain constant fraction rebalanced portfolio, with a small but solid mix of high and low volatility assets.
It's really hard to do much better than that.
----
Edit: upon re-reading your question, I think I misinterpreted it. Maybe you're just asking about how to find a small but solid mix of high and low volatility assets to have in your constant fraction rebalanced portfolio.
One start is reading Thomas Cover's 1991 paper on the Universal Portfolio. In order to understand that, however, you might need to back up a bit more into the history of E log X optimisation. The book "Kelly Investment Criterion for Capital Growth" is a collection of historic and modern papers on E log X, including Cover's. It's a very good book. Though it takes a bit of working through results practically to understand.
----
That will help you find a decent mix of high and low volatility. Specific numbers are no big deal, though. I would recommend at least 50 % low volatility, but some people (like Taleb) suggest up to 90 %.
The mix of high volatility is also not that sensitive to specific numbers. To avoid having to do difficult optimisation, just split evenly between whatever assets you want in there.
To avoid correlations, just look for things that seem like they're not cointegrated. There's no point in optimising this too hard either, because as I started out saying, when you need uncorrelated returns, they will be correlated.
I usually have been playing low-risk investments and want to change that, any articles or advice to make independent high risk bets that actually end up lowering overall risk?