Why do we ever get worse Mover values in comparisons?

[russell-taylor]

(done) 1xso, for example, has worse scores for reduce2 than reduce for some Movers.

Run with alt-A only in both Reduce and Reduce2.

Run with the version that checks all single-hydrogen-rotator angles and see if this persists. Check the ones that are worse to see why.

[russell-taylor]

The largest negative ones from alt-A only:

(done) SingleHydrogenRotator 0 'a' B ASER 100: -2.635; singleton

• (done) Why is the comparison score negative for this Mover for Reduce2 (-0.619)? There are no clashes shown in this residue when running probe2 on the output file. The score provided for this Mover within Reduce2 goes from 0.39 initial score to 1.61 final score so it should be positive. Rerunning both version using both alternates got rid of this discrepancy. In this case, Reduce2 is better.

(done) SingleHydrogenRotator 0 'a' B THR 114: -0.889; clique with AmideFlip B AASN 51

• This probably double counts the interaction between the two clique elements because a hydrogen from each interacts
• Reduce makes a hydrogen bond with a nearby O instead of making contact between hydrogens
• It did see the acceptor, so will have tried that contact

@todo NH3Rotator 0 'a' B LYS 67: -0.836; singleton

• Each had some clashes and some hydrogen bonds with a water and each had some contact with another arm

@todo: NH3Rotator 0 'a' B LYS 105: -0.573; singleton

• This score difference didn't have to do with the Mover, which was the same orientation for both.