@User (@User): Yes, negative numbers fall out of liberty tables due to the fact that the STA separates rise and fall delays from throughput delays; I wrote the Vesta STA program without researching the whole STA problem too deeply, and so I have never been quite sure what I'm supposed to do with negative timing values. For a single gate, it makes sense that with a slow rise time at the input that the output may transition before the input has reached, say, the midpoint of the transition. But if you don't limit that behavior somehow in the STA tool, you end up simulating non-causal circuits. . .