Anyone know Holley Commander 950's? programming?

It could be the "accelerator pump enrichment". On EFI it is done by looking at the change in TPS value. It triggers additional time for the injectors and decays it time. It should also clear when the throttle is returned to idle.

It is possible the the TPS is providing a noisy signal, or other noise is getting there. I have experience off idle hesitation on OEM vehicles related to high mileage TPS, where using an ohm meter did not detect fault, but a new TPS solved the problem.

I scope on the TPS output as you crack the throttle may give insight. Also monitoring the 5V supply that is used for sensors is important. It should be stable. The voltage is referenced by the analog to digital converter, so a problem there is widespread, for all analog measurements. Other adjustments to consider would be the TPS value at idle, there is often a small pre-load so there is not back lash, and the values that are set to deliver the enrichment.

Correctly implemented, throttle based acceleration enrichment in EFI is far superior to a carb. It works so well, I would not know how to improve.

I developed my own methods, but yours could be similar. My settings relate to a threshold throttle change, and fuel volume. The delay is not time but ignition intervals decay the value, this works well because at higher RPM, less enrichment is desired.

There may also be Apha-N adjustments. This is throttle based enrichments based on angle. For a MAP system, I typically zero the enrichments at RPMs below cruise. This for me is below 40% throttle, then above that (foot into it), provide additional additional ramp for power enrichment (AFR ~12). This provides good economy for normal driving, and enrichment when you need it.

All this changes if you have WBO2 and mixture tables.

If you do not have MAP, I have no idea how to make it work well.