As I alluded to at the beginning of the article, the PixArt PW3360 has long been considered the best mouse sensor on the market. It boasts one-to-one tracking and no built-in acceleration, angle snapping, tracking latency, or tendency to lose tracking. I've personally spent quite a large amount of time with the PW3360 in the Rival 700 and 500, and it's served me well and powered through every test I've thrown at it. Given that experience, I was skeptical when I first heard SteelSeries' claim that its new TrueMove3 sensor is the only sensor with true one-to-one tracking.
I looked further into this claim, and it turns out that that the PW3360 also has the true one-to-one tracking advertised in the TrueMove3. The real difference is that it's limited to a smaller CPI range. From 100 to 2,100 CPI, the PW3360 has pure one-to-one tracking. Beyond that range, the PW3360 has some jitter reduction, which means there is potential for some added latency beyond 2,100 CPI. The TrueMove3 extends the one-to-one tracking window out to 3,500 CPI. After 3,500 CPI, the TrueMove3 uses "advanced jitter reduction," which purportedly doesn't slow down response time. I have no idea whether or not the "advanced jitter reduction" is exclusive to the TrueMove3, or if it actually makes a noticeable difference.
Interestingly, the TrueMove3 actually has a lower CPI cap than the PW3360. The PW3360 can go all the way up to 16,000 CPI, while the TrueMove3 only goes out to 12,000. Honestly, this upper limit doesn't really matter because no one actually plays with CPI that high, as far as I know, but it does make me curious as to why the TrueMove3 can't reach CPI levels as high as the PW3360. If I were to make a wager, I'd guess the sensor is simply limited to 12,000 CPI because past a certain CPI, it's difficult to accurately track input or reduce jitter. Inflating the maximum sensitivity could potentially harm SteelSeries' claim that the TrueMove3 has one-to-one tracking.
Another intriguing question is whether or not it's actually a good thing to not have jitter reduction past 2,100 CPI. Jitter reduction is obviously needed past a certain CPI, because SteelSeries would simply do away with jitter reduction if it wasn't needed so they could claim its sensor has one-to-one tracking through the entire CPI range of the mouse. Maybe there's a reason the PW3360 has jitter reduction past 2,100. Was some sort of advancement actually made that allows the TrueMove3 to have accurate tracking out to 3,500 CPI without jitter reduction, or are SteelSeries and PixArt simply pushing the limit in order to advertise a larger one-to-one tracking range?
Unfortunately, I don't have the answers to these questions. However, I did put both the TrueMove3 and the PW3360 through various tests in order to determine whether the TrueMove3 is a good sensor and whether it's actually noticeably better than the PW3360. I ran all the comparison tests at both 1,000 CPI, my preferred CPI and within the true one-to-one tracking range of both mice, and 3,500 CPI, outside the PW3360's one-to-one tracking range and the upper limit of true one-to-one tracking for the TrueMove3.
I first loaded up MouseTester, collected mouse data while moving the mice horizontally across my mouse pad at various speeds, and plotted the data as xCounts versus time. The plots below show the raw counts from the sensors and moving averages of the counts. The dots and lines should mostly resemble smooth ups and downs reflecting the back and forth movement of the mice across the mouse pad. However, no sensor is completely perfect, and I can't move the mice perfectly smoothly across the pad, so there will be a few irregularities here and there.
I ran dozens of tests, but I can't show all the graphs at once, so I've picked out graphs from each set of tests that are representative of all the graphs from each batch and the consistency you should be able to expect from the mice.
The first set of tests was run at 1000 CPI and showed mostly smooth curves, like the ones seen above, no matter the speed at which I moved the mice across the mouse pad. This comes as no surprise since these tests were performed within the true one-to-one tracking range of both sensors.
The second set of tests run at 3,500 CPI showed similar results to the first batch: largely smooth curves regardless of speed. Interestingly, MouseTester doesn't show any noticeable difference between the TrueMove3 and the PW3360, even above the PW3360's supposed 2,100 CPI cap for true one-to-one tracking. In fact, in of all my tests, more so at 3,500 CPI, I actually ended up with more irregularities with the TrueMove3 than the PW3360. This may be hinting at an answer to the question about whether jitter reduction is good to have above a certain CPI, but I'm far from certain.
It's important to keep in mind that the time in these tests is counted in milliseconds. A few jumps and jagged edges in these tests don't really translate to breaks in smoothness noticeable to the user. For the most part, the raw counts and averages from both sensors stayed in smooth curves, which is what matters.
I also used MouseTester to confirm that the max polling rate of the TrueMove3 is indeed 1000Hz. The graph above shows that the sensor updates right around once every millisecond.
I continued the comparison tests in Counter Strike: Global Offensive with raw input on and mouse acceleration off. I had to set the in-game sensitivity as low as possible for the 3,500 CPI tests so the crosshair wasn't going wild with the slightest movements of the mouse. I first performed my standard one-to-one tracking test by placing two books at either end of my mouse pad and moving the mice horizontally back and forth between the books. I used bullet holes to make sure the the crosshair pointed at the same spot whenever pressed up against either book. I carried out this test in multiple in-game locations with varying distances between the walls and my character. The PW3360 and TrueMove3 both exhibited consistent one-to-one tracking at the two CPI levels tested.
I also couldn't get either sensor to spin out and lose tracking, no matter how fast I moved the mice across the pad or violently slammed them down at odd angles. Before I exited CS:GO, I played some Arms Race matches with both sensors at 3,500 CPI to see if I could feel any difference between the two sensors. Once I warmed up to it, 3,500 CPI was fairly playable with the sensitivity cranked all the way down. Despite playing above the PW3360's one-to-one tracking range, I wasn't able to discern any difference between the PW3360 and the TrueMove3.
Lastly, I ran a simple test to make sure the TrueMove3 doesn't have any built-in angle-snapping. Sure enough, just as I found with the PW3360, the TrueMove3 appears to be devoid of any innate angle-snapping. In addition to the various tests, I played a fair amount of Titanfall 2 and Overwatch with the Rival and Sensei 310. The TrueMove3 responded to my movements just as it should, and I never noticed anything unexpected or out of ordinary.
At the end of the day, as far as I can tell, the PW3360 and TrueMove3 are both excellent mouse sensors. Maybe there is some hardly noticeable difference between the PMW3360 and the TrueMove3 above 2,100 CPI that could give highly competitive gamers wielding a mouse with the TrueMove3 a minute edge over those with mice bearing the PWW3360. However, I don't see any evidence of such a difference. Either way, both sensors are fantastic, and the average gamer looking for a mouse with a reliable sensor shouldn't sweat it when picking between mice packing the PWW3360 and the TrueMove3.