I created a crude surface gravity map of Phobos using a Digital Elevation model (DEM) available on Astropedia$^1$. I assumed a uniform 1000 meter 3D grid of equal mass blocks to find the gravity field at the surface:
The reference frame of Phobos' surface is not inertial and thus the centrifugal force must be accounted for (and is accounted for). The 0° longitude point is (I believe) the sub-Martian point as Phobos is tidally locked to Mars.
Phobos is so lumpy and small (note the units above: µg's), the local "down" direction can vary significantly from the center of mass:
Jumping:
The unreliable source does indeed seem ... unreliable. Luckily it's not that hard to figure out how fast a person jumps if you know how high they can jump:
$v_{takeoff}^2=2a_{g}h$
For a half meter (50 cm) vertical jump (on Earth) the takeoff speed is $3.1 \frac{m}{s}$. Elite basketball players can get about 100 cm for a takeoff speed of $4.4 \frac{m}{s}$.
Simulating both of these scenarios in a simplified model of a spherical Phobos in Mars orbit shows (jumping in towards Mars):
So yes you could jump off of Phobos, if you were an elite basketball player...and could jump on Phobos. You gotta get down to get up.
I think the escape trajectory is better visualized as the phase error between the jumper and Phobos in Mars orbit:
The results are similar for jumping from the antipode (far side) of Phobos.
As for escape velocity; it gets difficult in this 3 body problem to even define escape velocity, let alone calculate it, so I won't try either.
- Willner, K., Oberst, J., Hussmann, H., Giese, B., Hoffman, H., Matz, K.-D., Roatsch, T., & Duxbury, T. (2010). Phobos control point network, rotation, and shape. Earth and Planetary Science Letters, 294(3–4), 541-546. https://doi.org/10.1016/j.epsl.2009.07.033