There's a lot going on here and a lot to unpack and so even though I add my own perspective, I can't help but put on my analytical hat and contribute that way.
First I'll refer to the article which immediately invokes the theoretical orientation of queer theory which I'm aware of and I have some reading in but I'm not that versed in it so I won't go deep in. But nonetheless the usage in academic scholarship and social theory is to effectively cross-interact the theoretical perspective with the topic in question and see what shakes out. Often this means making as best an argument one can with it despite the fact that the initial comparison or usage is arbitrary. We can sit here all day and take even just critical theoretical perspectives (queer theory, race theory, colonialist, feminist, Marxist, etc) and have them interact with games. And believe me, it's been done and will continue to do so if only because all forms of art are being examined similarly. We just don't tend to see it that often because they're buried in critical theory journals.
Anyway, the article effectively brings up the idea of queer relationship or trans-gender enactment by "playing" the opposite gender which I disagree with if only because there's a more fundamental mechanic at play. That of vicariousness and the interaction between player immersion, character, and design. Whether it be the dot in Adventure or Lee in the Walking Dead and everything in between, taking on the role of "player" first and foremost tends to inevitably produce some amount investment in the character being embodied. At the base level, keeping with the dot from Adventure, you can just be "you" or "player character" but with that comes the assumption and interpretation of environment, tasks, agency, meaning, and all the other things that come into play because you're playing a game. Otherwise the illusion wouldn't work and it's "just" a dot of color moving around other abstract boxes of color. Night Driver is another great example of this. You're supposed to be a car driving at night when graphically it's just falling pixels to create the illusion. However that simple tricks works to make the player take on that player role in this referential context (a driver at night and not just lights on a screen).
Now imagine blowing that out with characters that are more realistically rendered, have their own dialogue, voiced performance, and choices. And in this brings up the second half to my point which is the interaction with design. This very discussion actually hinges upon it by which I mean a game has to be designed to give the player the choice of gender or character creation to introduce this dynamic. What of games that only offer their definitive protagonist? And controlling for that, a game can give you choice of character but no control over the look or outcome of their look like a Bethesda RPG or MMO would. For instance, Costume Quest lets you pick which sibling to go with male or female but really there's no changing their look.
Therefore to consider the underlying issue of how players and player characters interact, and therefore how and why players choose which character to be, we have to take into account the different design scenarios which can produce different outcomes. For instance, how might the same player rationalize their experience when confronted with a game that offers no choice of character, choice of character but no customization, full customization but a mechanical gameplay experience (think more MMOs) and full customization but a more heavily story-based experience (Fallout 3, Mass Effect etc.). In other words, how deep does the "might as well hot ass" rationalization go across these outcomes and is that player also looking at story or characters in games different than one who sees that reasoning as irrelevant and/or off putting.
I know for me, I've come to realize that it really does depend on the game. And for full disclosure I too have found and continue to find the "might as well make it a hot lady ass" reason off putting to say the least. Generally speaking when given full customization I tend to make "me" aesthetically and that means a tanned skin bearded dude with black hair. Despite being vaguely Mediterranean/Turkish in appearance any attempts to make myself tend to produce Latino looking male characters. Which I'm actually totally happy with and adds to the experience; it's me but also different and also not an arbitrarily white male character either. Granted, with that comes the inherent bias of not stepping outside your bounds BUT in practice there have been important counter-examples. In Dragon Age II, I played the female character because I thought the actress was better done but with the fact that a number of major party members are female, I actually found quite a lot of value in the vicarious female/female bonds I made. And indeed it was compelling to project who I thought was an attractive male character for my player character to fall for etc. In the end, the choice of female had great value as a written character and as a means to explore that perspective as a male. I'd imagine that this won't be the last time a female character will inherently have added value as a chosen character.
Aside from the "hot ass" reasoning and my own above, there's one last perspective that is worth mentioning and that is the meta-representation reason which I take into account as well. Case-by-case many players would probably go along with their self-image biases but those choices are not done that way in a vacuum; instead they are done in consideration of games over time. At least for some players, there is a real consequence to the relative saturation of male characters in games and more so if one is not male. Therefore while a person may otherwise go with characters that reflect their own appearance and identity, there is a contingent who are going out of their way to pick female characters precisely because women are overall underrepresented as player characters in games. And for certain, individual motivations can be layered and can change over time. For instance, I might eschew the white dude character if it otherwise won't matter but if female character design is actually worse, choice is offered to customize, or there is a outstanding reason to go with a male character over another anyway, I'll do that.
With all that said there's arguably a whole other long post to make deconstructing the nature of viewing the chosen female character as having values exterior to character and player character immersion and as an attractive conception in and of itself. Namely an idea like moe and all manner of projected attractions in fandom of character works which need not extend purely to uncharacterized player avatars in games like WoW.