I think I found my mistake, it was a transfer error / typo... I propagated !c but wrote down a in psi1. Here is the corrected version:

Sadly, I still get the same error message (however I now somehow have 55 points??)

I also checked with the tool and got the following solution:

- ψ[0] = {s4,s6} = a & !c = CNF{!c ; a | c}
- ψ[1] = {s0,s2,s4,s6} = !c = CNF{!c}
- ψ[2] = {s0,s2,s3,s4,s6,s7} = !c | b & c = CNF{b | !c}
- ψ[3] = {s0,s2,s3,s4,s6,s7} = !c | b & c = CNF{b | !c}

I guess the tools just removes {!a,b,!c} from psi0,..,3 since {b,!c} already implies {!a,b,!c}.

Same for removing {b,!c} from psi1 ,2 because !c implies {b,!c}.