The executive summary answer is that the canned sims in CVCX used only the I18 indices, where my sims used the full set of HiOpt II Indices.

The Tarzan sims were the first time I published SCOREs which were generated using my software. In order to demonstrate that my results were not in left field, I also published HiOpt II+ASC SCOREs generated by CV for the same games alongside of my results. I did this by performing new sims myself using CVData+CVCX. The reason for that is that, as Norm just said in another thread, there are many many nuances and interactions which may not be obvious when configuring a sim and also when writing the software to perform the sim.

CVCX alone only gives you partial control over the sim parameters. By using CVData+CVCX, I was able to match the configurations of both CV and my own software as closely as possible. I also configured all of the sims to conform to the sim parameters of BJA3 Chapter 10, except for things like choice of indices used, since Don and I felt that it would not be fair to compare HiOpt II + ASC to Tarzan without full indices.

I'm looking at my records and, for the 5/6 S17 DAS game, 1-12 spread from the Tarzan sims. I see my SCORE of 44.19 and the CV SCORE of 44.93 for HiOpt II + ASC. The canned sim SCORE is lower because it used only the I18 indices.