It's true I made a generalized statement, and of course you can always make a poor hardware purchasing decision. Some purchases are dead right from the get go, and sometimes people just flat out overpay for things. However, acting within the right time frame, I think expecting 66% resale value for an ASIC isn't unrealistic.
However, the value of an ASIC is not "purely" determined by the BTC it generates because the value of an ASIC in terms if btc is mediated by fiat conversion rates. If I spend $1000 on an ASIC that makes 0.1 btc per month, it's a good decision if the value of btc is $10,000. It's better to receive 0.1 btc valued at $1000 than to receive 1 btc valued at $100. Of course, if the price of btc drops, then the profitability of the miners also drops. What I'm saying is that, given a reasonable purchase price, purchasing a miner with fiat may incur less loss than using that fiat to purchase btc.
Of course, the exchange rate also affects the dollar amount you can receive by reselling your equipment. What I meant was that there is no floor to the resale value of your mining equipment. With GPUs, people will always buy them to game, so you could always count on recouping at least that much. If bitcoin dies, ASICs become worthless.
Also, you would have to sell very quickly to get 66% of what you paid back when you sell your ASIC. If you assume that difficulty increases 20% forever, then you would need to sell after 3 weeks to get 66% back in a rational market. Obviously, difficulty isn't going to increase like that forever but it will continue that way for a long time and by then, you will be barely making anything so the calculations should result in a similar result.
So two general mathematical statements representing my point might be:
1) If btc price drops, buying btc < purchasing miner with fiat < do nothing
2) if btc price rises, do nothing < purchasing miner with fiat < buying btc
This is a possible scenario but unlikely. Most of the time, either buying the miner is better or it isn't regardless of the price of btc. You can estimate your return on the miner as the amount of btc mined and the amount of btc it will mine in the future, which should be equal to the resale value. If this amount of btc is greater than the amount you could have received by purchasing btc directly, then it will be the better option no matter what the exchange rate is.
The only way this changes when the btc price goes down is that the amount of btc you expect to mine in the future to increase. If btc crashes, less people will order miners and future difficulty won't be as high so you can mine more btc that are worth less. However, the effect of this is very small because most of the btc a miner will make in its lifetime will probably happen in the first few months since the difficulty will continue increasing for a long time.
Even if you end up profiting in terms of fiat because the price increases, that doesn't mean buying mining equipment is a valid choice. If you're offered $100 or $200 and you pick $100, then you will have made a bad decision even though you gain money. If you want to argue that buying mining hardware is a valid option, then either you think you can get such a great deal where a miner can produce more bitcoins than what you can get by buying directly or the small benefit in the event that the btc price decreases is worth the difference. However, I believe that the disadvantages of mining far outweigh this one small advantage.
1. You have to set up the miner and make sure it stays running. Most people run calculations assuming 100% uptime.
2. You receive btc over time rather than all at once. Let's say that some news comes out causing the price to drop. If you bought btc, you could sell them immediately. On the other hand, a large amount of your value is tied up in the resale value of your mining hardware and you can't sell it quickly.