I stumbled across this post while catching up on weekend reading. It got me thinking - my automated explorer currently has fixed limits on how many tickers a portfolio is allowed to own. But, I keep it limited to 25-30 at the max.
It makes me wonder what would happen if I dramatically expanded that to a substantial portion of whatever universe I'm looking at. The more tickers it models, the more compute the backtests take - but probably worth a gander.
"If you think a portfolio manager has skill, you want her to run a concentrated portfolio to increase the potential outperformance vs. a passive benchmark and get a faster signal if you were wrong, the portfolio manager doesn’t have skill even though you thought she had."
And we can then easily diversify the number of active managers held, to counter our overconfidence, without much chance of overlap and over-allocation.
I stumbled across this post while catching up on weekend reading. It got me thinking - my automated explorer currently has fixed limits on how many tickers a portfolio is allowed to own. But, I keep it limited to 25-30 at the max.
It makes me wonder what would happen if I dramatically expanded that to a substantial portion of whatever universe I'm looking at. The more tickers it models, the more compute the backtests take - but probably worth a gander.
Is this not the fundamental law of active management, reproven?
I like very much your point of Skill! Also I remember Buffet’s argument has always been that diversification is not the goal.
On the other side, recovering from losses is extremely difficult. Hence some form of diversification, be it 5 or 10 investments, is needed.
"If you think a portfolio manager has skill, you want her to run a concentrated portfolio to increase the potential outperformance vs. a passive benchmark and get a faster signal if you were wrong, the portfolio manager doesn’t have skill even though you thought she had."
And we can then easily diversify the number of active managers held, to counter our overconfidence, without much chance of overlap and over-allocation.
Prima!