This is a little confusing, but here you go.
When we invented the chip, we had three modes.
In 'short' we made the pulse repetition rate short, so in a giving amount of time we could send a lot of pulses, and get a lot of photon times. To range a bit farther we had Medium. Longer period between pulses, but you could range farther.
And of course 'Long' has the longest period between pulses - but you could get the longest range.
Now the tricky bit...
We invented 'extended range'.
During normal ranges we have a problem with 'Radar Aliasing' or wrap-around. This is when a target is beyond the max range, and the returning photons are associated with the wrong out-going pulse. It makes targets at 4.3 meters look like they are at 0.3 meters (when the wrap point is at 4M.) To prevent this we use two different pulse repetition intervals. (Call these A and B)
If range A is the same as range B, then we no issues. If they are different we know we have a wrap-around.
Then some clever folks realized that we could use the difference in range A and B results to not only detect the wrap issue, but to figure out where the target is.
So instead of a max 4M range, we could go out to 8M is some interesting corner cases - like when looking at a highly reflective projector screen.
But we notices that the best combination of timings for extended mode was "Medium'.
So, if you have extended mode turned on, use Medium.
If you want more signal, but are willing to give up distance to get it, use Short.
If you don't want extended mode, but need 4M, use 'Long'.
But for extend mode, Medium works best.
- john