Comparing HADDOCK scores with different restaints

Can scores obtained in different runs docking the same two proteins, with different restraints, be compared? Can one tell if one is better than the other of the four in the example below?

HADDOCK score: Buried surface area: Restraints violation energy:
-47.9 +/- 6.2 2047.5 +/- 64.4 70.2 +/- 46.86
-96.0 +/- 4.8 2258.5 +/- 137.0 79.4 +/- 3.01
-12.1 +/- 9.9 1578.3 +/- 104.4 26.1 +/- 15.14
-71.2 +/- 9.5 2395.7 +/- 273.3 47.8 +/- 29.39

You can indeed compare those scores.
What you could do however is to remove the restraint energy from the score (substract 10% of the restraint energy).
In your case there is a clear winner…