There are 5 key factors:
Feet:
An under-discussed topic that is certain to have a greater depth to it than people realize.
There are two materials used that I know of, smooth plastic and teflon. Teflon is preferred, although smooth plastic is not bad by any means. You can always add teflon mouse skates later if you wish. Teflon is a bit pricier than plastic obviously.
It is important for the feet to have round edges.
Weight:
I am of the opinion that the less weight the better. Weight adds to the friction. Also the extra force required to move the mouse dampens the variability of speeds that your hand/fingers/arm can create.
Mice with variable weights are silly imo because you would always want the least amount of weight, and having that extra function there wastes space on the mouse.
Shape:
Everyone knows how important this one is. A bad shape can cause hand strain, reduce mousing ability and take your focus away from the game.
Although a "one size fits all" policy is destined for failure, it is clear that some shapes are simply better than others. The 1998 apple imac "puck" mouse was hated for its shape, while the wmo shape is largely celebrated.
I would include the sensor position in this category. Most mice try to position the sensor directly below the knuckles on a palm grip and this seems natural. The sensor should be there, or further up the mouse closer to the front buttons.
The closer the sensor is to the front the more the mouse lends itself to finger gripping/aiming, the center more for palm gripping, and the back for arm gripping. Mice with sensors at the back of the mouse are absolutely horrid in my opinion. Also the sensor should absolutely _not_ be off center.
This mouse here suffers from both problems:
http://www.blogcdn.com/www.engadget....ue-track-1.jpg
Cord:
Braided cords fair slightly better than non. However even in non-braided varieties I have heard that some cords are superior to others in terms of malleability. Stiff cords can influence mouse movement, and provide resistance in certain directions. They also increase the necessity for a cord clip.
The wmo cord is notorious for being stiff and unpleasant. Some even strip some of the insulation off to improve its malleability. When it comes to a given mouse, your best bet is to simply ask around or look through various forum posts to see what the general opinion is on the cord.
Avoid cordless mice. The technology is still not good enough to be trusted for games.
Sensor:
The sensor is arguably the most important factor in computer mice, but also the most complicated.
I'd give it 8 of its own factors:
*Malfunction speed:
How fast the mouse can be moved before it either exhibits negative acceleration (due to skipped counts), or just plain goes haywire.
The first generation optical mice had very low malfunction speeds, and would make some kind of bizarre Z pattern if you exceeded it. Luckily what will mostly happen at the malfunction speed with modern mice is negative accel.
The speeds are usually listed in meters/sec. Low sensitivity players can hit over 1 meter/sec (roughly 6 inches in 150 ms) easily. In my opinion though 1.5 m/sec is much rarer, and anyone who regularly hits 2 m/s needs to get help for their cocaine problem before their mousepad catches fire.
You can see mice "vomiting" at their malfunction speeds here:
http://wikis.jp/interfacedevice/inde...aximumSpeed_en
Actually it is just the number of counts being registered dropping off as the sensor cannot pick up the intended number of counts past a certain speed.
For me personally, I almost never hit the malfunction speed of 1.5 m/s on my 500 hz wmo.
*Surface finickiness:
Some mice just plain don't like some surfaces. As to why I'm sure is quite technical and over my head.
When a sensor doesn't like a particular pad it will jitter and jump and/or just behave oddly. The most common example I hear for this is the abyssus disliking the qck pad. Although some people report no issues with this same combination. It should be fairly easy to determine what issues a particular mouse has with a given surface from forum posts and mouse reviews.
Many sensors cannot tolerate reflective surfaces, but that shouldn't be a problem on most pads.
*CPI:
Also known as "Counts per inch", or sometimes less accurately stated "dpi" for "Dots per inch".
It means exactly what is says, how many "pieces" the sensor can analyze in a given length.
The greater the number, the more smooth and coherent the input becomes. The question is whether or not increasing CPI is beneficial past a certain point that may be as low as 200.
Certainly for LG aiming the greater coherence seems to help as the fast adjustments can be made more smoothly.
At the same time there appears to be some drawbacks to "high" CPI values. Almost all of the top players used a CPI value at or less than 800, and most closer to 400. Why exactly are the feel of these CPI ranges so preferred?
I have two theories on this:
One has to do with what happens when the mouse slows to a stop. With higher cpi there is more mouse input between a point of motion and the point of rest. Picture two cars that are both going 40 mph and are going to stop at the same intersection. One car slowly decelerates to a stop while the other slams on the breaks. Both took the same distance to stop, stopped at exactly the same place, and experienced the same change in speed (from 40 mph to 0 mph).
However they have drastically different velocity curves. Less mouse counts effectively makes it like the car that slammed its breaks, since as far as the mouse is concerned your velocity dropped off faster than if it had measured more counts in between. This behavior is preferred in quake since it gives the mouse a "snappier" feel.
This seems to go along with my own experience with high CPI values. While I appreciate the extra smoothness, I find myself tending to give the mouse a more stringent grip to compensate for the lack of "snappiness".
The other theory I have is that the extra precision actually hinders aim since the sensor is more likely to pick up small jitters from your hand that a lower cpi setting would not. To give you an idea of the difference that can exist here, compare 400 cpi to 5000 cpi. At 400 cpi one mouse count is occuring every 63.5 micrometers, while at 5000 cpi one count happens every 5.08 micrometers. So the 5000 cpi setting is registering 12.5 times as many counts in a given distance.
To use a very rough analogy, it is like trying to balance on a unicycle with a shopping cart sized wheel vs. a bicycle tire. The smaller wheel might allow for much greater precision but it comes at a price since the tiniest gesture will influence the balance, even if you make the small wheel really stiff (analogeous to adjusting sensitivity to compensate for greater cpi). Too much of a good thing can be a bad thing.
*Prediction:
Also called "angle-snapping" and "drift control". It helps you to make straight lines by increasing the threshold needed to "count" as a deviation from a given movement path. The most widely used example is drawing a straight line in paint with the freehand tool. Without prediction, this is incredibly difficult even on a 400 cpi mouse, which makes sense. At 400 cpi even a ~63 micrometer deviation will "count", and I don't know anyone whose hand is that steady.
In games though it is thought that prediction is generally a bad thing. After all it is not allowing the true movement of your hand to be expressed to the computer.
Unfortunately a great many sensors have prediction built-in, such that it cannot be turned off or adjusted.
Some modern sensors have the option to turn prediction on/off, or even adjust it. (See "Features" below)
*Built-in accel:
For some reason, some sensors exhibit positive or negative acceleration that cannot be directly adjusted. As far as I know the main "pathogenesis" for this is counts being missed (neg accel), counts being counted more than once (positive accel), or the CPI setting varying with the speed. The steelseries Kinzu is notorious for suffering from built-in positive accel. Steelseries even came out with an embarrassing article on how they "meant" it that way to try and assuage negative feedback on the issue.
Some mice exhibit the problem only at certain CPI values. Almost all mice will exhibit negative acceleration past their malfunction speed.
Occasionally, although rarely, a firmware update can fix the problem.
Usually this information is easy to find as any mice with accel problems quickly gets a reputation as such. Even the sensors get a reputation, so any new mouse that comes out featuring that sensor will incur the same reputation.
*Lift-off distance:
Measured in millimeters, this is the distance the mouse can be raised from the surface and still track.
For this one I'd say the lower the better. When you lift the mouse to reposition it on the pad you do not want it to continue tracking.
To give you an idea, the wmo has a lift-off distance of ~3 mm which is adequately low. Could be lower if you ask me.
One crude way to measure this is with compact discs, as each one has a thickness of 1.2 mm (usually, apparently they can range from 1.1 to 1.5 but I don't believe it).
*Features:
Some sensors come with a boatload of features. The best example of this would be the Avago ADNS-9500 (xai/g9x).
It allows you to adjust the CPI in 1 cpi increments, adjust the amount of prediction (and switch off), add acceleration, change the polling rate in 1 hz increments, and even has a setting for "jitter control".
Some sensors are the complete opposite. The STMicroelectronics OS MLT 04 used in the wmo has no settings what so ever. I doubt it even supports firmware changes. It operates at 400 cpi.. and that's it.
Probably something between these two extremes is ideal, as in my opinion most of the 9500s options are not needed.
*Optical vs. Laser:
To be honest with everything that I've read and experienced, I don't think one is better than the other. I think it depends completely on the individual sensor being used.
The only difference I've found that seems to apply universally is the aperture. Laser mice tend to use a small opening that can collect dust/hairs and things, whereas opticals tend to have a larger opening, and as such are more resistant to debris. Some examples:
Laser:
Lachesis
http://www.techpowerup.com/reviews/R...ges/sensor.jpg
Xai
http://www.hardwarelook.com/images/r...ai/small/6.jpg
g9
http://www.everythingusb.com/images/...iew-bottom.jpg
Ikari Laser
http://www.cravingtech.com/blog/wp-c...-increment.jpg
Optical:
wmo
http://www.dansdata.com/images/morem...wmobase440.jpg
Ikari optical vs. ikari laser
http://www.dragonsteelmods.com/Image...e/ikario22.jpg
Kinzu
http://www.techpowerup.com/reviews/S...ges/bottom.jpg
CM Storm Spawn
http://www.hardwareheaven.com/review...spawn_base.jpg
Debris tend not to stick and just 'fall out' of the optical mice whereas they tend to stick in the laser ones.
My list of pros and their mice:
k1llsen - g9
garpy - deathadder
fazz - salmosa (wmo -> ie 3.0 -> salmosa)
av3k - salmosa (mx310 -> salmosa)
noctis - salmosa (frankenmouse: g1 built into an ms "trekker" ball mouse ->
salmosa)
cypher - abyssus (deathadder -> abyssus [450 dpi])
Cooller - abyssus (wmo [400 cpi] -> abyssus)
ZeRo4 - abyssus (wmo [400 cpi] -> abyssus)
spart1e - xai (deathadder -> xai)
calipt - xai or kinzu (deathadder -> roccat mouse -> xai/kinzu)
vamp1re - xai or kinzu
linkje - kinzuadder [400 cpi?] (wmo -> kinzu -> kinzuadder)
czm - kinzu
strenx - xai (?) (kinzu [800 cpi] -> xai)
dahang - kinzu [800 cpi] (ikari optical -> deathadder -> xai [1800 cpi] ->
kinzu)
chance - kinzu (salmosa -> white wmo -> kinzu)
DKT - DKT mouse (g9 -> ie 3.0 -> Zowie ec1 [500 cpi] -> DKT)
madix - imo 1.1a (ZoWie version)[400 cpi]
zsx - imo 1.1a [400 cpi]
stermy - ie 3.0 [400 cpi]
griffin - ie 3.0 [400 cpi]
z4muz - ie 3.0 [400 cpi]
v00 - ie 3.0 [400 cpi]
bodzo - wmo (Ikari laser -> wmo)[400 cpi]
fox - wmo [400 cpi]
jibo - wmo [400 cpi]
tox - wmo [400 cpi]
rapha - wmo (mx518 -> wmo -> xai -> wmo) [400 cpi]
Note: some of these may be out of date but should be rather close
Links:
Awesome list of mouse sensors/mice and their features:
https://spreadsheets.google.com/pub?...EE&w=100&h=650
Here was my review on the xai:
http://www.esreality.com/index.php?a=post&id=1848087
Guy on youtube who does pretty good reviews of mice:
http://www.youtube.com/user/Ramla777
ZoWie Mico review:
http://www.youtube.com/watch?v=7aNYX26Gu8M
Girl doing something odd with a CM Storm Spawn:
http://www.youtube.com/watch?v=R3zRNevJyns&t=1m42s
. Highest I go to is 1500dpi, but 1100dpi is good enough for most games.
I had an MX518 before which was great, but it had path correction, which I wanted to get rid of. The WMO is also so much lighter, don't think I could go back to the 518 now.
