Is any acceptable range of Lambda in slope/w?

In stability analyses, sometimes very weird slip surface come up for the critical slip surface, especially when we perform "block search". Then if we plot the FS vs Lambda, very possibly the converged Lambda value is less than say minus 0.2. Do we have any empirical range of Lambda, which is corresponding to the acceptable solution? Otherwise, the block search approach is basically useless。

BTW, what are " root finder" and "Linear Search". Try to understand them, nowhere we can find the explanations.

Thanks

Shan