On warrior vs scout starts and worker stealing:
When I stated that scouts can't steal workers, I had flexibility in mind. The situations where opposing scouts can beat you to huts are generally the same situations you have a decent shot at sniping their workers (not always but often enough), and barring a powerful tech from the hut, a stolen worker is superior. Of course, if you're isolated you can just make a scout later.
The other major problem with hunting is that resources it allows are rarely in an initial BFC, making it limit worker utility early depending on the other tech (combo of mysticism/hunting can really hurt sometimes).
On choking:
Generally I shake my head at players who consider a choke "cheap". Sirlin.net discussions aside, the AI gets 90876340987 bonuses at high levels to begin with. If one is to label chokes cheap, then you might as well label war bribes/religious manipulation cheap, along with a ton of other strategies. I prefer NOT playing with my hands tied, honestly.
I'd also like to point out that choking in the wrong situations can be extremely counter productive. Here's a couple reasons it's more balanced than people think:
1. The AI you choke will tend to hate you. No OB, no early trade routes with them, no tech trades, etc.
2. How many AIs do you intend to choke? It gets costly to choke multiple AI's, but starting with difficulties higher than noble (and especially monarch on), choking one AI tends to let a different computer REX out a large share of that land. You might eventually take the choked city, at the expense of creating a fairly monstrous AI nearby. For this reason, the power of a choke and its benefit to the human player varies considerably depending on the situation.
3. Unless you're playing without barbs, the worker stolen has to be escorted home typically. Virusmonster noted this as a problem in his really high score walkthrough and disabled them. Basically, if you steal the worker and want it back at your capitol ASAP, it may be hard to choke the AI.
4. While the AI will spam archers like no tomorrow, it WILL eventually send a settler with escorted archers out, meaning to wipe the sucker out you'll probably have to take multiple cities.
5. Realistically, to make use of a choke to actually wipe out the AI will require the construction tech, which means you're going to have to reinforce the choke and balance the excess costs of units outside the border with research, all the while being handicapped a bit since an otherwise potentially peaceful neighbor is at war with you.
These things make the effectiveness of a choke highly situational.
It's so ridiculous that it's considered cheap. "The AI isn't programmed to handle it". Hah. The AI isn't "programmed to handle" human tactics in general in civ 4, that's why it gets such insane bonuses to compensate.
What's next, calling parking your siege/CR SoD outside of a city instead of in, inviting the AI stack to capture it and then wiping it out with ease cheap? The AI certainly doesn't seem to be programmed to handle a tactic that can potentially allow a 4-1 kill ratio at tech parity. That can be a lot more game breaking than some choke, too. At this rate you could make a case that few people go through a game without doing something "cheap".
That's why I hate hearing that word in games, at all, ever. Some things might not be balanced correctly, but I'd say having prats or quechas is a much greater advantage than the ones afforded via chokes in many situations, making the argument that "choking" is cheap but playing inca/rome somehow isn't seem ridiculous.
Edit: All this and I never even mentioned some other points about chokes: your starting warrior has to live (which is not a guarantee, even if you're careful), the enemy has to be reasonably close, and you have to actually bump into enough animals/xp hut to hit lvl 2. That's a lot of ifs.
The AI also has to have enough forests so that you can actually reach the worker when it moves. This is not true for 100% of starts, and in some rare scenarios you may not even be able to scout this ahead of time.
Finally, while a woodsman II warrior is decent against archers, it's not 100%. On high levels the AI has access to some decent bonuses and will spam archers with alarming ease, improvements or not (and pillaging flatland resources next to the AI capitol will generally get you killed). If all you're using to choke is that one warrior, the AI will eventually attack it with 3 archers or so, and woodsman II still leaves a warrior behind an archer until you count fortification bonuses. 2 archers will win almost every time. Certain AI's will attack the choke more aggressively than others (aka shaka and such), meaning you'll have to sell out to hold it. As I already mentioned, this comes with some potential expenses...
For these reasons, I steal workers fairly often, but only rarely commit to a choke.