Personally, I think it’s about how impresonal modern matchmaking is.
You’re only ever playing against “enemies” and “enemies” should be “hated with the hot passion of a burning sun”. And if you lose, you’re never at fault, because your teammates sabotaged you!
People don’t have to maintain cordial relationships, because they will never meet their teammates or opponents again.
Compare that to stuff that works using servers, where each team is made up of the same pool of people from one round to the next. People actually make friends with each other, friend or foe, and have more fun as a result.
If you play enough, pure random chance will eventually get you a game that feels like a fair fight.
But quite often, video game matchmaking systems will fail to accurately estimate player skill correctly, creating teams where one will utterly demolish the other.
The same player isn’t going to perform identically every session, and accounting for every possible weapon or character/class they might play, potential synergies with teammates, or potential advantages/disadvantages in matchups against any given opponents…
It all makes for a literally infinite number of variables, all of which must be accounted for.
The correct way to get interesting matches, imo, is to make it semi-random, and not try to have all the players on both teams be exactly the same skill level. Rather, put players on both teams from a range of estimated skill levels. This way both teams have weaker links for the other team to potentially exploit, and both teams have strong players which will try to stop that.
Instead, the system should just enforce common sense stuff, like not pitting someone who is literally playing for the very first time, against a team with someone who is 2000 hours in, and hence might straight up deny the new guy a chance to play at all.
I should know. I literally wrote THE team balancer for titanfall 2 community servers. For a time it even used the Tone online database of player stats, to know how to balance players that had never played on a given server before.
I was genuinely shocked how good the resulting games were. All I did was take the completely random players that decide to join a server, and simply figured out a slightly smarter way than other balancer scripts at the time, to divide them into two teams that are close enough to equal.
Changing voltages and fan curves is super situational. And depends on how much you value noise over performance.
That said, I undervolted and underclocked the i7 cpu on my G501 gaming laptop back in the day.
This helped a ton, because the heatsink between the discrete GTX 660M and the CPU, shared a heatpipe. The CPU would only throttle at 90, while the GPU would throttle at something like 75. This meant that because it was basically always hotter, heat from the CPU would conduct via the heatpipe INTO THE GPU, causing it to always thermal throttle, and be unable to be cooled. Because even though it was maxing out and trying to cool down by throttling, the CPU would just keep going because for it the temps were fine. So it would keep pumping heat into the heatsinks and heatpipes, which would then keep the GPU hot, too.
Undervolting the CPU allowed it and the GPU to run at closer to same temps, raising FPS by way of allowing the GPU to actually run a full tilt, even though the CPU was then significantly slower.