Different results between SweeD 4.0.0 and 3.3.1

[SC-Duan]

Hi all,
I run SweeD 4.0.0 for my vcf data, and I set grid as (chr. length / 1000), and used -folded parameter, however, I got ~99% Likelihood values with 0.000000e+00.

Position Likelihood Alpha StartPos EndPos
487.0000 0.000000e+00 1.200000e+03 487.0000 836.0000
1487.0199 0.000000e+00 1.052632e-01 1373.0000 1781.0000
2487.0398 0.000000e+00 4.095563e-02 2194.0000 2942.0000
3487.0597 0.000000e+00 2.575107e-02 3021.0000 4955.0000
4487.0796 0.000000e+00 2.823530e-02 3021.0000 4912.0000
5487.0995 0.000000e+00 5.882353e-02 5283.0000 6097.0000
6487.1194 0.000000e+00 1.333334e+00 6394.0000 6496.0000
7487.1393 0.000000e+00 1.791045e-01 7420.0000 7671.0000

When I chose different grid value as (chr. length / 100), (chr. length / 10), the results were same.

And I also run SweeD 3.3.1 for the same vcf data with parameter -folded and set grid as (chr. length / 1000), this result looks right.

Position Likelihood Alpha
487.0000 1.645027e-01 3.638509e+01
1487.0199 2.143608e-02 2.788293e-02
2487.0398 1.517468e-02 1.460734e-02
3487.0597 6.375422e-03 6.360513e-03
4487.0796 0.000000e+00 2.823530e-02
5487.0995 1.315065e-02 1.271961e-02
6487.1194 0.000000e+00 1.333334e+00
7487.1393 1.105182e-01 1.291514e-02
8487.1592 1.783558e-01 2.946162e-01

What caused the different results and how can I modified?

More, for the result of SweeD 3.3.1, I can get the 5% Likelihood values as candidate, and how can I get the region on chromosome (StartPos to EndPos)?

Thanks!