1 stream vs 6 stream tests the consistency of your connection. It sounds a little strange, but the more you load your connection with traffic, the more efficient it becomes, as the bottleneck will always be your internet connection vs the speed of your LAN. However this means your router will be queuing up loads of packets to pass through the bottleneck and how efficiently it manages this really matters.
Normally, with 'default' settings from an ISP router, it can lead to an undesired effect called 'bufferbloat', where the router's buffers in both your router and the ISP get full, which reduces latency and general efficiency. In a simple way, you can run a speed test while also running a ping in a console and you will see the ping times go up dramatically while the speed test is running. The TBB speed tester (which along with the tester at
https://www.dslreports.com/speedtest) are the two best testers on the Internet short of using special tools. When you do a test on TBB, it will have a button for 'Analysis', which will grade your line from A to F, and the same from DSL Reports will give you a rating of 'Bufferbloat'.
To mitigate this, with a decent router (OpenWrt, pfSense and a few commercial ones like Asus) will allow you to set up something called QoS (Quality of Service), which is usually based on a technology called 'fq_codel' (Fair Queue, Controlled Delay). Sometimes they are just called 'Smart Queues' on commercial routers. fq_codel is proven to be one of the most efficient traffic marshalling algorithms, by ensuring the maximum amount of traffic can go through the bottleneck, but delaying other packets so it's never overloaded.
Using this, your x1 stream and x6 stream will look broadly similar, and most importantly, your latency will stay low, so when your wife looks at a web page, it doesn't slow her down, when doing DNS lookups and her web page will still be very snappy while your big download only drops by .1mbps or so at the same time.
Here's a speed test of my line:
https://www.thinkbroadband.com/speedtest/15851378587...
My line is quite troublesome, as I'm about 650m from my cabinet, and have a 17.4db attenuation. Theoretically my line could sync at 68.6mbps and 20mbps, but I deliberately limit it to 58/18 because Openreach's DLM will introduce interleaving if I run it 'peddle to the metal', which introduces latency and increasing mine from 5-7ms to over 24ms. 58mbps is more than enough to watch 4K streams, and really I couldn't give a damn about an extra 1MB/s on my downloads of 7.5MB/s. On top of that, I then run fq_codel on pfSense which moderates the traffic at that 58/18mbps limit and thus the 'analysis' on the TBB speed test shows a quality of 0.1 or grade 'A', and the idle, down and up bars are generally pretty flat, but you can also see this in the graph of how the x1 and x6 line are also pretty much the same to see the effect of what QoS does.