Both portfolios use the same AI ranking factor, trading system, universe, and period. I understand that live execution may cause small differences, but why is the live strategy down 9% while the simulation is down only 1.3%? What could explain such a large gap? Also, Why is the predictor selecting different stocks? (Same predictor)
Looking at the graphs the use of the mega-cap weighted S&P 500 as benchmark is not providing useful information. Look for a benchmark that more usefully reflects what you actually hold. this is a “least worse” exercise and not a “best fit” exercise.
The return pattern looks to be similar, but with a 20%+ haircut versus the sim. Which implies a too optimistic value for slippage like 0.25% when it should be 2.00%. Looking at the bid-ask spread of what you are buying should provide some information.
Turnover is also higher than in the sim with 4 versus 2 trades which adds drag and a butterfly effect on what is actually owned.
Don’t be surprised that when you get the sim to be closer to your actuals over this period that it is back to the drawing board.
Cheers,
Rich
This year a tiny timing can lead to dramatic differences in realized trades. I experience the same thing live recently with daily rebalance. A lot of breakout trades fail recently and in this earnings season a lot of AIFactor favorites crash hard post earnings. From the shape I suspect this is a microcap system?
Check and compare the realized trades. Sims use Saturday ranks and the start date defines the trade day for weekly rebalance. Live strats use actual daily ranks, so ranks get updated the day after earnings for the ticker, NOT like in the simulation on saturdays. The only truly comparable sim/live setup would be weekly on mondays for both.
Usually live performance can be even a bit better than expected since you get the new rank info daily not weekly and you always trade on freah info. I don’t know what’s going on this year but many US strategies seem to fail hard recently. It looks like a bit more stocks, a bit looser rank tolerance, a bit more conservative universe and your US return is not 0% but suddenly 30% YTD.
Thank you all for your comments and responses. I really benefited from your input, and I was able to solve the issue.