I don't know how this is calculated on the A330 (rated on A320), but on the A320 this is the way the system is designed. If you are high on profile the intercept point is valid for a descend with half speedbrakes. So without speed brakes the intercept point will move forward. Not sure if its the same on the A330 but from your description it seems to be...

Pond Crossing
 

WRT Uplinkers reply earlier,
and just as a question, would the selection of "Constant Mach" in the FMGC not be more normal on a NAT Track rather than pulling a Selected one? Also WRT the original posters comment on the thrust maybe hunting from idle to climb does the cost index make any difference to this....?

Anyone thinking 777 is speed stable in turbulence (even light turbulence) has probably never flown it!:rolleyes:

Airbus is fine. Each product has it's pros and cons, though I accept one manufacturer's product is much more like the basic machine people would have trained on, and is therefore likely to be more intuitive in the early days. After a while Airbus becomes moderately intuitive too, though some features will always present a challenge. Normally connected with tactile/visual monitoring of flight control/auto thrust inputs!

Like I said. Both products are fine.

777 isn't perfect by any stretch of the imagination but it is better than the 330.

744, now there's a good machine, rides bumps very nicely:ok:

One technique that some people do that I've seen cause big power fluctuations
is a very gradual VS climb while maneuvering in a block altitude. Like, VS +200 FPM or such. Usually imperceptibly smoothe, but if the winds are variable through that altitude block, the power can get pretty active. This would certainly not explain feeling surges like that for most of the flight, though.