Noticed some problems at S1 state tracking optimization: 1. As soon as S1 (dominated 22211000) falls below the old S0 (dominated 22220000), its states are beginning to "run away" - the weight of 22211000 gradually decreases, and 22220000 - increases. Can this be avoided? Freeze the state 22220000 or remove it? 2. After a certain point the states (S1 and S0) reversed (despite state tracking!). See attached out. Maybe it's effect of the first problem.
This message contains the 221 kb attachment
[ p1_pt90mco2_3.rar ]
![[ Previous ]](/ceilidh-icons/ceil_prv.gif){kind=icon}
![[ Next ]](/ceilidh-icons/ceil_nxt.gif){kind=icon}
![[ Index ]](/ceilidh-icons/ceil_ind.gif){kind=icon}
Wed Mar 23 '11 9:06pm![[ Reply ]](/ceilidh-icons/ceil_rep.gif){kind=icon}
![[ Edit ]](/ceilidh-icons/ceil_edt.gif){kind=icon}
![[ Delete ]](/ceilidh-icons/ceil_del.gif){kind=icon}
This message read 784 times