mudmanc4

At first glance the latency 'dip' appears as network congestion during the test, some type of QOS, or if this is happening on all tests, could be power fluctuations within the first line node. Or any number of other issues actually.
 
As far as speedtest, it's not showing the latency because it discards such findings. Why is anyone's best guess. But great for ISP's since it will not show issues such as you have pointed out in testmy.net results.

Yes! 
I hit that thing up from time to time , just to reminisce. Lots of good times. 
 
Best thing is , when i start getting itchy about having bandwidth flukes, I look back and see what we were all bragging about at the time , puts things into perspective  

You TiP shows a dead zone more or less, might not be the proper terminology , most the way through the testing period. Actually degrading through the test.

At some point, it would appear that packet filtering or inspection is in heavy use. Once the same data is transmit repeatedly, it gets the attention of this at this point 'anomaly' , and something begins to inspect more close. I'm only speculating here.

If you run several tests different times of the day and different days, take notice if this is a constant.

Thanks for your thoughts. I should have marked this solved.

The correct way to do this would be to use capitol " -P " when specifying a port in this manner.