Intentionally stressing the candidate in an irrelevant way is a dick move that tells you nothing.
It's not irrelevant. It allows you to judge a person's reaction to a sudden and inconvenient situation. Surely you've come across such a situation in your years of work. Contrary to popular belief, work is not one continuous process of milk and cookies. Every so often an event occurs which necessitates rapidly changing direction to complete a task or goal.
Seeing how people react to a stressful situation is a fair test. Yes, you will lose good candidates using this method, but you also lose good candidates because of your bias as well as whatever filtering process HR uses to get you the list of candidates.
That belief is such a sad waste of time and energy. I've been on both sides of the interview table for over a decade now and I've never encountered anything in that time that suggests you might be on to something.
For starters, high-stress interviews can very easily out your organization as a bad place to work to the candidate and filter out the people you actually want while filtering in people who are only proving they're good at your interview process.
I think about 7-8 years ago, Facebook wanted me to write a freaking calendar app just for the opportunity to stand before wave after wave of teams of climbers looking to filter out anybody who wasn't a Stanford comp. sci grad. Meanwhile, FB the site at the time, was having mad cross-browser issues and had been for over a year and there was a huge shortage of developers with a ton of skill/knowledge on the front end at the level that I was at. That told me a story about Facebook behind the scenes that I didn't like. Even if I got through the live-fire exercises and they only happened to hit the on the algos that I was aware of (and have never used in a production environment before or since), I didn't love the idea of what I was going to find on the other side of that rainbow. So I had a choice. Play the FB lottery ticket for what sounded like a dubious-at-best work experience with great compensation/benefits or talk to like 7 reasonable companies in the same time frame with a lot less stress and fatigue incurred.
And who would put up with this process? A minority might be people who don't actually have any other choices but somehow ended up getting picked up by a recruiter and will just waste your time. But most likely, people who are so in love with the idea of working for Facebook, they'd eat shit for Facebook. Young people who spend their nights honing all of those interview skills that often have no correlation with practical experience skills. That's what you want? A bunch of shit-eating badge-earners rather than than people who love what they do enough that quality of work environment trumps the best possible compensation/benefits and brand-recognition? How many "people who actually quit Google" articles do you need to read to figure out what you're doing to yourselves?
So...
* Whiteboard? Huge nope. I've been typing like a pro for 33 years because I couldn't handwrite for shit. It's not even possible. It would take me like a minute to write a legible line. I've stopped asking people to accommodate me on that because even when they grudgingly accept they still end up being pricks about it at the interview. You want it on a whiteboard, you're probably being an asshole. Completely unnecessary. Also we're not REALLY engineers. Get over it.
* 8-10 hour "test" projects that will actually take the better part of a week to complete because you have to polish every last bit to avoid being filtered based on somebody's favorite linter preferences? Big nope. Also, WTF do you need 8-10 hours of work for, even if that was realistic timeframe? What are you telling the candidate about how much respect you have for people's time? Can they probably expect the same of you in the workplace? From what I've heard, yes. If you're convinced this is such a great idea, at least tally the number of candidates who reject you when they find out what's on the menu and ask yourself who those people probably are. False negatives are worth it for the best candidates, my eye. You're shooting yourself in the ass and you're probably too silly to know it because the same process put YOU there.
* Gauntlet of pass/fail test interviews. People will find the dumbest reasons to filter out candidates when you handle things this way. Stop filtering. This is a matching process. Establish the base line of competence/experience that you need and then cut it out with the code-marathon stuff. It's ridiculous.
* Non-trivial pair-programming. Oh my god nope. I've actually warmed to pair programming recently but don't force candidates to code as you watch. You're a stranger to them. No trust to make the occasional quickly-discovered derp without getting judged has been established.
It's just not that hard. Give them something small to write or refactor or the option of bringing in recent work to show so you can discuss. OR have them bring in something old they might be proud of but would do very differently now for discussion. Have them solve very typical day-to-day problems in person that should fall within the realm of the experience they're supposed to have. Ask them about challenging problems they've had to solve and how they went about solving them. Most importantly, ask questions with an aim of establishing whether they actually enjoy the work. Because if they don't, they might not suck today but they will most likely suck within 3 years when they drag their feet every new thing they have to learn.
Also, what's with all the posts with the character entity issues? Are you guys typing your responses up in Word 2003 or something?