I asked NASA via Twitter about it hoping they answer it during the Orbital Sciences/Cygnus-2 Post-Launch News Conference, but Stephen Clark from SpaceflightNow.com beat me to it and asked Frank L. Culbertson, Jr., Vice President and General Manager of Orbital's Advanced Programs Group, this very same question over the phone. It can be seen in this video roughly 1 hour 19 minutes into it. Here's a transcript:
Stephen Clark, SpaceflightNow.com: At the launch video, it looks like there's a little bit of a tilt of the vehicle coming off the pad.
Is that a normal maneuver or something unexpected?
Frank Culbertson, Orbital Sciences Corp.: You asked about what some of us call "the wiggle of the pad", others call it The
Baumgartner Maneuver, but it was a little bit more pronounced today,
or maybe we could just see it a bit better.
As the rocket lifts off, it actually moves the tail of the rocket
closer to the structure, so that it then flies away from the
structure. That's to protect the tail and the surrounding hydraulic
equipment and the electrical cables from the blast of being too close
to the rocket. It's only a degree or so, but it is sometimes visible,
depending on the viewing angle.
Yes, it is a heart-stopper when you see it and you don't expect it,
but it was all as it should have been.
This so called Baumgartner Maneuver, named after Paul Baumgartner, Master Engineer at Orbital Sciences, is additionally explained by antonioe in one of the NASASpaceflight.com Forum posts, also regarding Antares:
What you saw MIGHT be the "Baumgartner Maneuver" - an initial pitch
AWAY from the TEL makes the TAIL of the rocket actually get CLOSER to
the TEL ... but then the resulting (small) lateral acceleration
carries the rocket AWAY from it ... this maneuver had to be very
carefully designed, a trade between the initial tail movement towards
the obstacle which is the inevitable result of the pitch, and the
I was told (by Paul B. himself) that the pitch-away and the resulting
tail-towards-TEL motion would be so small that they would not be
perceptible ... maybe he was not counting on telephoto lenses and HD
video (and sharp NASASpaceflight subscribers...)
And therein is described what I believe was the reason for the somewhat more pronounced effect of this maneuver this time around, as captured by the broadcast cameras - their position relative to the launchpad must have been slightly different than during previous Antares launches from Wallops and they captured this Baumgartner Maneuver better.
As for the second part of my question, what is the maximum lateral drift that would still permit the launch vehicle to navigate through the cabling between the four lightning towers, it seems I assumed wrong and the Mid-Atlantic Regional Spaceport Launch Pad 0-A doesn't use any cables between them at all, like e.g. those at the Cape Canaveral Air Force Station Space Launch Complex 40 do. This becomes more apparent by looking closer to some of the photographs of the Launch Pad 0-A, for example these posted on Spaceflight101.com page to Cygnus Orb-2 Mission Updates.
So, judging by the photographs, the maximum tilt of the launch vehicle during liftoff could be several degrees more, and it would still be able to clear all the launchpad installations (mainly TEL or Transporter/Erector/Launcher and the umbilical cabling).