Subject: Re: Another possible screen
Not sure the effort and time is worth it when ETF XLG top 50 only costs about 25 basis points.

Always with the negative waves, Moriarty. ;-)

From a backtest at testfol.io, Jan 2006 to Jan 2024, starting with $10,000.

Name      Ending    CAGR      MaxDD     Vol'ty    Sortino
           Value
XLG       $75,298   11.22%    -52.40%   18.95%    0.81
VFINX     $65,528   10.40%    -55.26%   19.52%    0.74
QQQ       $143,689  15.06%    -53.41%   21.88%    0.97
AJG       $163,374  15.84%    -47.89%   22.84%    0.99
BRK-B     $76,438   11.30%    -53.86%   21.54%    0.8

Top20     $66,848   10.52%    -52.89%      <<< My backtest 
XLG       $76,080   11.3%     -48.3%       <<< My backtest

Neat thing about testfol.io is that you can specify the range of dates instead of being limited to the last 10 or 5 or 3. These are the comparisons from the first year that all the above were available on Jan 1.
My excel spreadsheet backtest shows essentially the same results as testfolio, which confirms that it is valid.

So I'd agree with your assessment -- might as well go with XLG.

VFINX is the S&P500. BRK and XLG both did (slightly) better. CAGR 1% better.

However, both QQQ and AJG trounced them. At a cost of slighty worse volatility.

FWIW, I also backtested using GTT timing, and got the typical results. Return was a little lower but MaXDD and volatility and Sortino ratio were all much better.