Rosewill RK-9000 v2 Single-Button Speed Test
There are two metrics that I measured to determine response from a single button: how quick after a keypress can I input a second one; and how many times can I press a key in a given timeframe? These tests are obviously very dependent on switch type, but there are also slight differences between keyboards of the same switch due to the keycap used and other factors.
For each of the following tests I provided error bars. Given the random nature of these tests, my actual performance with each keyboard in each test would be somewhere inside those ranges about 68.3% of the time. That of course assumes that my arm never got tired or anything.
To determine how quickly I could mash a button I found a simple game online and played it over and over. Each keyboard was tested ten times and the score was averaged. I also calculated the standard deviation to quantify how much my results vary.
The Cherry MX Blue switches were fairly identical and slightly behind all others. The Cherry MX Red performed best at this event and was quite consistent in its results with a score of 46.2 actions in 5 seconds (9.24 Hz). Despite five of the six data points being whole numbers, they were averaged over ten data points.
The Cherry MX Black is possibly the biggest story here. Its performance suffered below the Cherry MX Brown switch, but had a substantially large range in scores. I believe this is due to the Black switches being very quick when you attempt to ride the actuation point but very slow if you must bottom out the switch due to its firmness. It is possible that when you really get to know the keyboard that you may be able to ride the activation point more efficiently. This means that the low average might be a factor of my lack of skill rather than equipment.
To determine how much latency is between button presses I developed a little Java applet to time the duration between button presses to the millisecond. Each keyboard was tested ten times and the score was averaged. I also calculated the standard deviation to quantify how my results vary.
The Cherry MX Blue switches surprisingly cleaned house with this test. It often feels very laggy when I attempt to dodge in Unreal Tournament 3 and yet when measured it performs even slightly better than the MX Red keys. I cannot really explain that.