It's a choice between a higher throughput into the funnel, or a higher conversion rate. Neither is necessarily wrong.
^ My one problem with the upfront credit card system is the <$1 test transaction which is usually applied and then instantly refunded. This makes the anti-fraud systems at my bank go into meltdown and puts a block on my credit card until they can confirm the transaction with me.
So I can't see why you would sacrifice more potential customers (higher funnel throughput) for an artificially higher conversion rate.
I mean the conversion rate is just a metric. A conversion rate of 5% of 1000 is better than 10% of 400. The goal must surely be, at least as long as the cost of serving a tyre-kicker is almost nothing, to get as many to try as possible.
Yours and the OP's points were well-made. I think it's ultimately empirical which works best for your product. We haven't quite gotten around to testing that, so for the time being, we're sticking with upfront cc because it's simpler.