MultiShot - it takes time to fire the 2nd shot (it’s 0.1 seconds after the first) and this reduces your speed by 12.5% (best guess); I haven’t timed Triple Multishot so my guesstimate is that it will reduce by 25%
These percentages are additive - e.g. if you have 8.5% from Nimble Hands, 12% from Battle Pass, 25% from a Big Attack Speed up, 70% from Scorcher, 100% from Sustained Fire and -12.5% from MultiShot, then your speed would increase by +8.5% + 12% + 25% + 70% +100% - 12.5% = 203% (so a Plasma Cannon firing at 1 shot/second would now shoot 3.03 shots per second)
Note: I’m reverse-engineering this from “60 fps” screen recordings (looking for the frame when the shot appears on screen) and that includes errors from the recording being VFR (variable frame rate), so my the observations are 1% to 7% off from the predictions (the error is greater for fast guns, where miscounting 1 frame has a bigger impact)
What do you mean by “0.1 seconds after the first” for multishot? Is it that
It extends the time between shots by 0.1 seconds, and that’s what causes the speed decrease, or
There’s a flat 12.5% speed decrease, and the second shot just appears 0.1 seconds after the first
I assume it’s the second?
Like, if it’s the first, then the slowdown would be more severe for faster guns. In fact, with stress shot + sustained fire + autogun, the shots/second could go over 10. Then, if multishot adds 0.1 seconds per shot, then it would actually slow down the shots/second.
But also, if it’s the second, then in the stress shot + sustained fire + autogun scenario, then second shot from multishot would come after the next actual shot.
It’s been a long time since I analysed this and my memory is a bit rusty; but my guess is that the reason there’s a slowdown (of ~12.5%) is to have time for the 2nd shot to fire (which is a fixed 0.1 second delay) before the next actual shot, in most scenarios - but in extreme cases (like your example) it could indeed be that the next actual shot is due before the multishot.
Unfortunately at those kind of speeds analysis from video recordings just isn’t possible for me (maybe a phone with 240 fps recording could do it) so I don’t know what would happen - hopefully it just interleaves the shots (e.g. Shot 1 First Shot, Shot 2 First Shot, Shot 1 Multi, Shot 2 Multi etc) but it could be that some shots get lost.
I’ve also seen claims from time to time that Attack Speed is capped, but I haven’t been able to test that either (for the same limitation of fps)