Accuracy may seem like a helpful stat, but without a controlled environment, it is useless to compare accuracies from different groups.
For example, if I play BR starts against Marine bots I can get about 60s and 70s (occasionally break 80 as well).
Against ODST bots with BR starts, it drops to 40s and 50s (with breaking 60 on good matches).
Also, if I do Marine bots BR starts and no shields, i.e. SWAT, my accuracy gets labeled as 20s to low 30s. This is because first bullet kills on a headshot and the next two are counted as miss.
Only way we could really get good accuracy comparisons on inputs is a controlled experiment.
For example, 1000 controller plays and 1000 M&K players agree to individually boot up custom games, play a game against 8 Marine bots with no shields and commando start. We may get a better picture.
The no shield and commando is to allow one hit headshots with a more forgiving aim than the sniper rifle, and without as many possible misses from burst fire.
You’d have to record the data from your own PVP matches to get a good idea of how good you are in the real world. Play ten matches of ranked, add up your accuracies then divide and you’ll get a true average. Or just use Halo Tracker.
It’s not even good in equal settings purely because some players play different roles. A player who rotates around the map and flanks will naturally have better accuracy than someone who has to engage the enemy a lot to prevent pushes over longer distances, applying pressure. Both are equally important to have in the team.
Some players have mad trigger discipline and others will absolutely pepper the life out of everything, pre-firing at every opportunity. The pros have always had a large disparity in general BR and accuracy ability. Some literally can’t snipe to save their lives but they’d still run rings around most players with their game sense.
Check out Halo Tracker then. There’s a lot more data there to extrapolate and quantize. Even considering different play styles. You’ll be able to check everything on a match by match basis. You can even track specific players
This was just a suggestion for the whole debate of AA, and you make a good point of another factor that would need to be controlled to truly say if controller is better than M&K.
There is a reason they have so many other stats now, to show more what you mention and that fact will always be high accuracy does not mean awesome player.
Yeah my bad, I just realised what you were going for. That’s actually a very good idea although I still think accuracy will be heavily swayed by people’s playstyles.
Something I’ve noticed playing on mouse is that the ranked lobbies are far less experienced in general. This may partially affect the accuracy too.
Oh yeah for sure. Think going for a large sample pool of players could even out the play styles potentially, but would need a lot of people to opt into a controlled experiment.
It would be interesting too to plot and see where distributions lay for each input as well.
Still a flawed experiment. Each weapon has separate aim assist values so we need to compare across the board.
Let’s use snipers instead, that way when keyboard players get 2-3x more accurate than controllers we can claim aim assist needs buffing. That’s how this works right?
Hahaha, yeah. It would need an experiment for each weapon even.
Just goes to show trying to collect the data to even prove this is an issue is a nightmare.
In the end, we will likely not have definitive data as there are too many variables to control and becomes a question of which weapon to use as a baseline tester.