Optimized to two decimals.
Wrong by eighty points.
91 . 9
Optimizer weights when stock A is forecast 1 point ahead
vs
9 . 91
Same two stocks . the forecasts swap by 1 point
A friend showed me his spreadsheet on Sunday. Solver had picked his weights. 47.32 pct in one stock, to two decimals. Looked like science. I asked one question. What happens if a return forecast is off by a single point. We nudged one cell. The whole portfolio changed sides.