Art of War 3 has a large variety of units (30 in total as of the writing of this article) and each player has their personal preferences. As there is no detailed official data provided by Gear Games, a survey was created instead to find out more about it ( Download Link (pdf) ). So unlike other wiki articles, which are meant to be as objective as possible, this article will be very subjective as it is based on player opinions. Therefore it is important to note that just because a unit is very highly/lowly rated that one will not always win/lose when using it.
On top of that, any details on how the data was created will be skipped here, but can be looked up in the document above.
One important note thing to mention here, is that all units could be voted as F tier (lowest) to S tier (highest), but the worst any unit got was a C+ (Mammoth), which is why the graph above only shows the relative strength of all units compared to each other, looking at the C+ to S tier range and cutting off everything below.
A completely average unit would here lie on the big red line in the middle and everything within reasonable distance from the average within the red dotted lines, allowing one to categorize units into 4 categories depending where in the graph they appear:
Over Powered
Above Average
Below Average
Under Powered
It needs to be said though that one has to be careful to now simply say, that everything in group 1 has to be nerfed and everything in group 4 has to be buffed, because it is not that simple.
Possible factors of why the data could be inaccurate:
Accessibility: For a unit to matter it first needs to get to the battlefield, so if a unit is unreachable for some players, like for example Thor and Albatross at low ranks, where resource production is much lower than on higher ranks, then those players will find barely any use for it, no matter how strong it is. Not to mention that for obvious reasons, higher tier units should outperform lower tier units, otherwise there would be no incentive for players to tech up to higher Headquarters levels during a battle to access higher tiered units.
Controversy: If something was voted as C tier, was it because most people voted it as C tier, or was it because everyone voted it as E and A tier, averaging on C. This heavily points toward a unit being very situational, which can be either a good or bad thing. The best example here are Fire Assaults, which are currently the strongest anti-infantry unit in the game, but very weak against everything else, making them very strong vs only 1 type of unit. So if a player meets that type of unit a lot, they will find it useful, and if not, more or less useless.
Control: Another factor to consider is why a certain unit is so strong or weak, because while some units can be used to more or less their full potential without any micro whatsoever (mainly infantry), other units like Vertex and Hawk depend a lot on micro management, which in extreme cases can multiply their strength as much as ten times of what it would be without.
Obviously there are more factors which will be more or less important which can further falsify the data, but the main point which should be kept in mind, is that there is always a gap between subjective and objective statistics, but they will at least provide an overview, making them better than having no data at all.