@joanneloveslife When I heard Virgin Media was finally ditching the Puma chipsets for the Broadcom chipset I was very excited and hence my reason to join Virgin Media. I'd like to call out again for a Hub 5 instead of Hub 3.
The Puma flaws were fixed in new chip production about four years ago, the problems with chips that had already shipped were resolved by a very effective workaround. I had a Hub 3 with the old silicon and the latency problem, I saw the issue before and after the fix, I know for a fact the fix worked (I was extremely surprised that it worked, but all credit to Intel's technicians it did). As I've now gone from a Hub 3 to a Hub 5, I know for a fact that the minimum latency is measurably higher for a Hub 5, the peak latency is every bit as variable as, so the net effect is much the same or worse, and on top of that I've had hour long disconnections almost every other day since the Hub 5 was connected. Hopefully that's a teething issue or even unrelated to the Hub 5, but it's not good.
Now, by all means continue demanding a Hub 5 if you're prepared to take on an unproven product that has no performance benefits over a Hub 3. That is a valid choice - I took the Hub 5 on because I was prepared to give it a chance and help VM create a test base to iron things out. But don't base your reasoning on years out of date news, and tin-foil-hat science from the bufferbloat loons on Reddit. If you're a fully paid up bufferbloater, then DOCSIS is not the technology for you, nor is VM the ISP for you, as VM operate at higher contention levels than Openreach, and even without utilisation problems the latency impacts of traffic queuing at the CMTS are often visible.
You should also bear in mind that I'm running in modem mode, so I've managed to avoid all the quirks and flaws in the Hub 5's wifi capabilities. Those who are relying on the Hub 5's wifi can expect even more quirks. Eventually I think the Hub 5 will be VM's best ever DOCSIS hub, but it isn't at the moment, and the bug fix priority list will be blatant faults and compatibility problems, not optimising latency performance for gamers.
Here's another few of my BQMs. The first two are the first Friday and Saturday of March, with a Hub 5:


And these were the first Friday and Saturday of February with a Hub 3:


The evidence is in front of you. This isn't opinion, it's fact, measured by Thinkbroadband's test systems. Would you still like to tell me that simply because it has a Broadcom chipset the Hub 5 offers better performance than the Hub 3?