Oddity during delta bed calibration

Just something odd I've noticed  when I hit auto cal for my delta, one of the points is almost always probed much slower than the others  it looks like it's about half the speed. It's random, too. Not always the same point. The piezo is sensitive enough it's not an issue, it still triggers fine. Just wondering if anyone else has encountered this? It was doing it on 1.18.1 firmware. I just upgraded to 1.20beta6, and it did it again.
Here's my bed.g file (I could clean it up some, it has some leftover stuff from experimenting with various zprobes and settings):
M558 P1 I0 F230 X0 Y0 Z0 T3000 R1 ; F120
G31 X0 Y0 Z0.125 P400M350 X32 Y32 Z32
G28
G1 Z5 F4500
G1 X0.00 Y119.00 Z5 F4500
G4 S1G30 P0 X0.00 Y119.00 Z99999 H0
G30 P1 X59.50 Y103.06 Z99999 H0
G30 P2 X103.06 Y59.50 Z99999 H0
G30 P3 X119.00 Y0.00 Z99999 H0
G30 P4 X103.06 Y59.50 Z99999 H0
G30 P5 X59.50 Y103.06 Z99999 H0
G30 P6 X0.00 Y119.00 Z99999 H0
G30 P7 X59.50 Y103.06 Z99999 H0
G30 P8 X103.06 Y59.50 Z99999 H0
G30 P9 X119.00 Y0.00 Z99999 H0
G30 P10 X103.06 Y59.50 Z99999 H0
G30 P11 X59.50 Y103.06 Z99999 H0
G30 P12 X0.00 Y59.50 Z99999 H0
G30 P13 X51.53 Y29.75 Z99999 H0
G30 P14 X51.53 Y29.75 Z99999 H0
G30 P15 X0 Y0 Z99999 S6G28
M350 X16 Y16 Z16 E16:16 I1
G28 
You have selected analog mode (P1) in your M558 command. In this mode, when the Z probe reading reaches 90% of the value specified in the P parameter, it will slow down probing to 1/3 speed. Looks like you are sometimes getting a reading of 360399 at the start of probing. So either switch to digital mode (P5), or increase the P parameter, or adjust the sensitivity of the probe.

Awesome, thanks David! Glad there's an explanation for how random it seemed.

Makes sense of this thanks, I've noticed this too. But since it didn't seem to cause any trouble just ignored it.