and/or does anyone have a critique about it?
The strategy is to buy stocks in proportion of their (expected) yield. So Let’s say I want to buy stock A yielding 10%*, stock B at 8% and stock C at 6%. I would buy them in the ratio 10:8:6 making my portfolio
A 41.67%
B 33.33%
C 25%
Still reasonably diversified but with a higher percentage in the better performing stocks.
Yield could be thought of as dividends + price appreciation for this exercise.
For vocabulary, yield is not the right word, that means dividend expressed as a percentage.
Total return means price appreciation + dividends.
First of all, payment of a dividend says nothing per se about a stock’s expected return. If a company has some cash, it can either pay it out as a dividend, in which case the stock price drops by the amount of the dividend (total return zero); or it can retain the cash as an asset that will continue to be reflected in the share price. No shareholder value is generated by paying out cash as a dividend per se. In practice, “growth” stocks tend to pay low or zero dividends because they reinvest profits in R&D; steady established businesses tend to pay out profits as dividends.
In any event, future total return is the big unknown, so I don’t really see how you can use it as an input to your portfolio allocation. How do you plan to estimate it? Broker research? Stock pickers notoriously underperform the market. And if you did hypothetically have access to highly accurate forecasts of stock performance, why wouldn’t you just spread your investments among the stocks with the greatest forecast return, subject to including a large enough number of stocks for adequate diversity?
One related concept I’ve encountered is the Kelly criterion. Basically, the premise is that when you find a bet with a positive expected value, you should take it… but it’s foolish to put all of your money on a single bet, because no matter how high the expected value, it’s still possible that you’ll lose. The Kelly criterion says that the percentage of your bankroll you should be willing to spend on a bet is equal to the percentage of its expected value: Thus, for instance, if a bet has a +5% expected value, you should spend 5% of your bankroll on that bet.
I suspect that this is only an approximation to what Kelly actually proposed, good for low expected value, because it fails at high expected value. It’s possible in principle for a bet’s expected value to reach or exceed 100%, but it’s still not wise to bet your entire bankroll on any bet, no matter how good the expected value.
Yes, Kelly does make sense. In practice, it wouldn’t have occurred to me to use it to design a stock portfolio, because we don’t generally have such precise long-term expected return forecasts. For short term trading purposes (especially something like M&A risk arbitrage) where you may have a much clearer anticipated risk/return profile, it’s certainly applicable.
It’s certainly a standard approach to portfolio asset allocation to use historical volatility to measure risk.
But the idea is that if I use the Kelly criterion to invest in say 5 stock then how could I weight my portfolio to higher expected returns while maintaining a reasonable balance. I get Riemann’s objection but that we are of course gambling which means we need to bet with imperfect information. So what if we assume that the expected returns stated are reasonable but understood that reality may be different (hence the diversification).
With all financial investments it’s logical to spread your risk.
No matter how good a single investment looks (here’s one that paid 10% for 17 years running :smack:), there’s always the risk that you lose all your money.
The wise investor goes for several of shares, property, bonds and even a gamble (here’s 'a really good gamble. 
)
This is the key point IMO, once we get past the terminology issue of the term ‘yield’ used, apparently, to refer to what should be called ‘expected total return’.
It all revolves around how accurate this information is. And there’s good reason to believe there’s no such thing as accurate information as to differences in the total expected return of various stocks independent of differences in risk.
For example, it stands to reason that if the exact same company with the same prospects finances itself 100% with stock issuance that stock will have a lower expected return than if finances itself 80% with bonds paying interest lower than the expected return on the firm’s assets (as would typically be so). All the upside is the latter case is packed into the remaining 20% financed by stock issuance. However just as obviously it’s much more likely the value of the firm’s assets will decline to 80% of the original value (wiping out the shareholders of the leveraged version of the company) than that they decline to 0% of what they originally were (and wipe out shareholders of the unleveraged version). But the investor could replicate the higher returning higher risk version of the stock by borrowing money to buy the unleveraged version. Hence a stock with higher expected return due simply to more leverage isn’t a ‘better stock’. The higher expected return has to come from some special sauce besides more risk, and the stock picker has to see that when not many other people do or else the price would already reflect it.
Tough, good evidence says basically impossible, unless you’re a quite exceptional stock analyst, in which case you should probably do that for a living with Other People’s Money. Of else you can do it for fun with a small portion of your portfolio but then it doesn’t matter how many stocks because it’s a small % of portfolio for fun.
But to summarize and reiterate, you must IMO focus more on the assumption you can find higher returning stocks (corrected for risk), before taking that dubious assumption as a given and asking how to implement a strategy based on it. That assumption is very unlikely to be valid for any given person.
I don’t think you have enough information to do this. If you did, it would certainly be a winning strategy to invest more of your portfolio in those stocks with the highest returns, but it’s kind of like having a baseball strategy of getting extra base hits with runners on base. It’s a lot harder done than said.