Debunking "if we reduce block times to 1 min then services will just require 10x more confirmations"

I used to also agree with the statement in OP until I was corrected. The correction makes sense -- the needed math and explanation was published in Section 11 of the white paper in 2009 but it can be counterintuitive.

The chance of block reversal is proportional to the hashpower the attacker has and the number of blocks being reversed, and has nothing to do with the total amount of work being performed.

If the attacker has 10% of the total hashpower, then they have a ~20% probability of being able to reverse one block, but only a 0.00012% chance of reversing 10 blocks. Again, refer to section 11.

The above calculations hold as long as orphan risk is negligibly low. It's true that as block times are reduced, and as blocks become larger, that orphan risk increases. However, technologies like xthinner mostly eliminate the orphan risk from large blocks.

Note that reducing interblock time does more than just reduce the time to settlement. It is also equivalent to a block size increase. 10 1-min blocks carry 10X as many transactions as 1 10-min block.

For further reading: https://blog.ethereum.org/2015/09/14/on-slow-and-fast-block-times/

Edit: I'm not here to necessarily argue in favor of a reduction in interblock times, but simply pointing out that we cannot have this discussion when so many people misunderstand the basic facts involved. Let's clear up the misconceptions, and then we can have a good discussion.