Amplitude is the rotation in degrees of the balance wheel as it oscillates. Depending on movement type and the age of the watch this will range from around 250° to 310° in the flat positions for a freshly serviced watch in good condition. Amplitude will fall by 15°-30° in the hanging positions, that's normal and it's due to the increase in friction at the balance pivots as they run horizontally in the jewels rather than vertically on one end stone. Friction in the pallet pivots also increases for the same reasons.
Rate is self-explanatory, it's measured in 5 positions (dial-up, dial down, 3 ,6 and 9 hanging positions).
Beat error is expressed in milliseconds. For beat error to be zero the balance has to be swinging exactly the same amount in each direction, for the watch to be considered correctly 'in beat' the error should be below 0.5 in each position. For a modern watch with adjustable beat arm and minimal wear to the escapement parts a value of 0 to 0.2 should be easily achievable. This is much harder to set on old watches, that's the point I made referring to Omega's relaunch of the 321 movement.
What's important is the spread between the rate figures in each position, the lower the better. For a watch to meet COSC the spread has to be no greater than 10 secs between highest and lowest, that doesn`t guarantee it'll meet COSC but if it's greater it definitely won't. A freshly serviced watch should be giving good amplitude figures which are top end of the range for that movement.
Have to confess, I don`t know how the machine calculates the other values! I can work some out but DVH and Di aren't obvious. It's customary to check the change in rate and amplitude after 12 or 24 hrs and also the change in the spread of rate values, the lower the values the better, that's what separates the best watches from the also-rans.
I use a far more basic Timegrapher and write the numbers down, can`t afford the top-notch machine that Duncan at Genesis uses. The expensive machines do everything automatically, moving the watch from one position to another automatically then spewing out a printout.
Hope this helps. These figures confirm that the watch is in excellent condition and running as it should. With rate figures varying by figures of +1 and +2 the measuring errors of the machine almost become significant! This is as good as it gets.
Edit: what the figures don't tell you is exactly how the watch will perform in everyday wear, that depends on wear pattern and the user.