In Michael A. Hiltzik's wonderful book Dealers of Lightning: Xerox PARC and the Dawn of the Computer Age, he described a weekly meeting that took place at the famous Xerox PARC research laboratory called "Dealer", after a popular book at the time called "Beat the Dealer". There was a speaker selected each week, called "the dealer", who was responsible for setting the topic for the week as well as the rules for the debate. The dealer would propose an idea and then try to defend it against a room full of the most critical, and brilliant scientists and engineers in the world at the time, who would basically attempt to rip it to shreds.
You may think that they took peer review to extremes, but when you consider some of their achievements at the time1, they may have been onto something:
So how does the dealer win?
The first step in convincing your technical peers of your argument is preparation. Your technical argument must be compelling, as opinion alone rarely cuts it: do your research so you can back your argument up with facts.
When you are presenting an opinion or a fact, prefix it as such: be honest with your audience, that will engender respect.
Appeal to the laziness of your audience, many of them will correctly equate laziness to efficiency. A high barrier to entry will hurt the adoption of your ideas by your colleagues, make their path to your desired destination easy.
Know your audience. Do not make fun of the work of your predecessors, as they or their friends may still be in the audience, and annoying them will automatically make them resist your ideas, even if they privately agree with you.
Finally there is only one thing that trumps a working demo, and that is a working prototype that is easily reused and extended by your colleagues. Again this plays into the strategy of providing a path of least resistance for the naturally lazy.
If you cannot defend your position, you are not in a secure place. You should try to think of critical questions or analysis as unit tests of your ideas: the more that fail, the more likely you need to reconsider your argument. Do not be too proud to walk away from a weak argument: your peers have done you a massive favour by picking holes in it.
During a debate, do not be afraid to say "I don't know that yet, but I will follow it up", and do follow it up and come back with an answer, even after a few days have passed. Do not bullshit, a smart audience will see right through it.
Finally, the troll in the room should be undermined mercilessly (there is always one, and they will identify themselves via their questions). Win the room by overcoming the loudest dissenting voices.
You have to ask yourself "Why do I want to win this argument?". If you sell your ideas successfully, you have influence. Why you want influence is up to you, we all have our agendas. For me, I try to make my working life easier by getting my peers to adopt simple, practical, and robust standards that make all of our working lives easier. That appeals to my laziness.