The problem is you criticize a prediction model without having the parameters, so of course you'll get to the wrong conclusions.
The fact is it's you who made wrong conclusions

I did not, we went for a risk / reward play based on the info we had. I don’t think it’s that hard to understand.
Yesterday, I sent you my proposal for a statistical method that reduces the 2^70 key space for puzzle 71 by 7.22%. Did you see it?