AKA "The Art Of Controversy" AKA "The Art Of Always Being Right" AKA "Dialectic Eristic". Here's a pretty fun, illustrated version of the text, in actual Troll terms]. Here's an audiobook.

EDIT: In this article I adopt a bit of a Devil's Advocate attitude. I'm not entirely convinced of what I'm suggesting, but I'll try to give it my all to make it look at least worth considering. I might get carried away at some points and overtly relish the villainy like a mad Britannian prince, which is unsightly, and, more importantly, unwarranted, so please forgive that. I'll leave those elements in, so that this is a Self Demonstrating Article.

So, the rationale is as follows: sometimes you get in an argument with someone. You're not quite sure you're right. You're not quite sure he's right. Even if you play fair, there's no guarantee it's the truth that'll come out. A few hours later, you could think up of an argument that would have saved your cause, you just failed to think of it during the discussion itself. And usually it's not just a matter of finding the truth.

First, it's a matter of "being right". If you want to clash intellects, there's no more violent, crude, intimate way than this. When you're proven wrong in a discussion, especially in public and in a way that makes you look like an idiot, your ego could get hit hard. Not to mention your status. Back when this book was written, people killed themselves, and each other, over this stuff.

Second, beside your own pride and life, there might be stuff bigger than yourself riding on this. You just can't afford to stick to the truth, or to give up just because the other side has better arguments. You gotta win, in the eyes of the public, no matter what.

This book makes a fairly good job of singling out different tricks to bullshit your way into winning an argument. Or at least stall for time and take your opponent off-balance and distract them while you think of something legitimate to say. Let's review a non-comprehensive list of the tricks he proposes (the cartoon site and the full text are much more adequate, having one or many examples per case and being very eloquently phrased by the writer himself.

Let's classify them by blocks:

• Attacks to the opponent's morale:
  1. being an openly unfair and insolent prick just to piss them off. 
  2. claiming vicotry in an authoritative, assertive voice, despite the argument going against you
  3. interruptions and diversions, derailing
  4. if they're angry about some particular argument, rub it against their faces until they lose it: it's probably a weak point in their defense. Same thing if they are being evasive.
  5. invoking arguments that use obscure sources and are hard to check
  6. appeal to consequences: show them that defending their argument means going against their own interests in a way they didn't think of. They'll drop it like a hot potato.
  7. confuse the hell out of your opponent through nonsensical pompous speech that sounds authoritative
  8. personal attacks and insults
• Strawmanning (making the opponent say something they didn't actually say, amking their position look worse than it actually is): then attacking the strawman. Also, making your position look better than it is.
  
  1. By overgeneralization and slippery slopes
  2. Exploiting double meanings, homonyms, unclear definitions
  3. Using loaded words, buzzwords, and guilt by association.
  4. Using false dichotomies ("with us or against us") and other false syllogisms to extract outrageous things from your opponent's proposition that weren't even there in the first place
• Checkmate: the cleaner sort of tactic, and the most humiliating, they rely on making the opponent sabotage themselves. The favored type of method of the Ace Attorney games as well as the more heroic court drama.
  
  1. Getting the opponent to admit to your premises (or even the premises of your premises) one after the other, without letting them know that they lead to your conclusion all along. Then draw it. It's safer to draw it yourself, but it's more fun to make *them* draw it and then watch their expressions as the absolute horror of what they've just done dawns on them. Mwa ha ha. One way of doing this is by using questions, Socrates-style, possibly in disorder so they stay off-balance. An especially fun way to do this is getting them to say no to propositions you fake needing him to agree on, then submitting the antithesis of what they just negated, which they'll have no choice but to admit.
  2. Using one counterexample to blow up an entire generalization, which crumbles like a house of cards. Especially effective if the counterexample is a Black Swan your opponent isn't familiar with.
  3. Using their very arguments against their thesis, mostly by pointing out implications they missed. Especially fun if the argument is false in the first place, but is part of the core dogma of whatever cult, party, or group the opponent pledges allegiance to.
  4. Angering the opponent into strawmanning their own position through exaggeration by way of exasperated reaction to your incessant bugging. ("YOU CAN'T HANDLE THE TRUTH")
• Jumping to conclusions:
  1. Making the opponent admit to the premises, then making the conclusion yourself, sometimes by generalizing his admissions to specific cases as admissions to a general truth
  2. Begging the question
  3. Using a faulty proof to reject the wole proposition
• Just plain make do with bullshit:
  1. Your opponent uses a sophistic nonsense argument. Instead of taking the time of exposing it for what it is, you just counter it with bullshit of your own
  2. Appeal to authority rather than reason.
  3. "It applies in theory, but not in practice". If the theory does not apply, it means it is wrong, period.
• Escapes and getaways, sometimes of the cowardly sort:
  1. Petitio Principi: refusing to admit an argument that would immediately lead to the opponent's desired conclusion by exploiting the fact that your opponent and audience didn't notice that little step and confused the premise with the conclusion and claiming it begs the question. One of the more subtle of the bunch.
  2. Defense by subtle distinction: if your opponent has blasted part of your proposition, claim to have been misunderstood, and squeeze and narrow the original proposition down to something your opponent didn't get to disprove. Save face, salvage what you can.

I'm just surprised Schopenhauer isn't an Internet idol by this point. I'm also pleasantly surprised at how our discussions avert most of this stuff, most of the time. Then again, our motivations are different from the usual, are they not? But what about our relation to the general public? Suppose one of us accepted an invitation at O'Reilly's? What about convincing people to donate when there just isn't time to convince them of how important our cause is or how we are the right people to carry it out (not to mention we're not quite consensual or certain on either ourselves)?

Spartans were famed for their laconic way of communicating. In fact, the term was named after them. It was an actual course in their education: teachers would mock them and provoke them, and the kids would be punished harshly unless they could respond quickly, forcefully and wittily. I think we should train ourselves on this. There is a time and a place for careful deconstruction of the opponent's arguments, and careful weighting of what is right and wrong. There's another for trusting in the heuristics you're following and acting on them now. Sometimes you just have to win, and worry about the truth later. So we should learn to identify when exactly the gloves should come off, and learn how to take them off quickly, so that we are never taken off-guard. If the very existence of humanity is riding on this project, I think a little verbal swashbuckling is the *least* we can allow ourselves in terms of consequentially moral leeway.

Not that just sticking to the truth is entirely ineffective, but opponents aren't always as malleable as the one in that example, we're not all as smart and witty as Eliezer, and sometimes the inferential distances are just too huge not to resort to Dan Browned, Conviction By Counterfactual Clue, or Lies To Children for the sake of expediency (there's an entire rule in Schopenhauer's book dedicated to the case of debating of technical matters before an untrained public, and he provides a really good example, to boot).

This article suggests that learning about, and perhaps embracing the dark arts may be a useful if not outright necessary necessary means to achieve our goals. The author, on the other hand, isn't so sure. However, at the very least, I think we should know about this stuff, if only in a Defense Against The Dark Arts way, and make and study a list of similar, more contemporary works that would give us a better results-to-time-investment ratio in learning these tricks and others, and, more importantly, their counters.

BTW, Robert Greene's books, despite being rather unscientific, are very promising in that regard. Their advice is fairly useless if you want to apply it, but once you've gone through all the contents (and there's a lot of stories there) you'll be on guard against practically anything: it's really hard to beat the Epic Fails he lists there, which are all the more epic because usually they involve smart, perceptive, strong, powerful people, and they all still fall for the exact same tricks, over and over again. If only because they are fascinating narrative anthologies, and a very fun intellectual read, and we are very much in favor of fun and intellectualism, right?

Also, for those that have followed this article from the start, notice how the successive rewrites make it a self-demonstration of the "defense by subtle distinction" rule. Whether its use here was legitimate or not is left to the reader.

EDIT: As usual TVTropes never fails to pleasantly surprise. Here is their wittily written, fairly comprehensive list of fallacies: they called it You Fail Logic Forever. Remebember that fallacies are just part of the Dark Arts of Winning Debates, and a very dangerous bluff if your opponent calls you out on them, second only to counterfactual arguments.