InsuranceGuy wrote:I was backtesting what you are proposing and got some strange results as the optimal allocation based on the backtest was to not use the HBPP at all and just to use the GEM allocation strategy for 100% of your money. I thought about it more and convinced myself that the GEM system generated enough excess returns with little enough added volatility for that to make sense. I took it a step further and I was able to use some of the other successful modifications I have made to the HBPP to enhance the GEM strategy returns further but ultimately I prefer using a REIT as my 5th asset as opposed to International Stocks.
Anyways, the thing that bothered me is that I when I compared my GEM results using the asset class returns I have previously mined to those published on the optimum momentum site and there are some significant discrepancies in recent history and even more the further you go back. As a word of caution I do think the GEM momentum system is a good system, but I am apprehensive about the returns shown on Antonacci's site.
Hmm, that would be troubling. My tests mostly matched Antonacci’s month for month since 1988, but I see differences before that, which I chalk up to evidently backfilling different data for the S&P 500 total returns with my numbers rolled from Shiller. I also don’t have aggregate bonds data and use 5 year Treasuries instead, which affects all the risk off months. And I don’t trust my synthetic Tbills all that much, which affects whether it’s in stocks or not any given month. All in all I didn’t expect an exact match and I didn’t get it
But it was enough to test the concepts and see how it works.
Are you seeing discrepancies not explained by any of the above?
I found it surprising and disappointing when Antonacci removed the page that listed what asset GEM was in each month historically. That was there when I built my spreadsheet and it helped a lot to check my logic.