I've definitely heard that argument you might be right ... Mining has the initial hardware costs, plus power all year, etc... plus difficulty and block reward decreases... There's really too many variables over the course of years, but in my overly simplified math example (with tons of missing variables i'm sure), I think an argument could be made that at a certain increase in cost , mining becomes more profitable:
(let me know if I'm wrong, I'm wanting to learn and share like most everyone else here) :
-BUY/HOLD -Buy $4200 of a coin at $0.5 each, if cost per coin doesn't change I own $4200 of coin
*in 1 year at 3x cost it's $12,600 I hold ,
*in 1 year if it 5x cost it's $21,000 I hold,
*and if it's 10x cost it's $42000. Solid.
-MINE- Spending $4200 ($2400 piece of hardware, plus $1800 in power for 1 yr /$150 a month) , and the average of coins mined daily over the year is 30 coins a day...
*in 1 year at same cost I've mined $5475 (30 coins a day, $0.5 each coin) minus cost of hardware and power = $1275 , not good
*in 1 year at 3x , mined total $16,425 - $4200 hardware / power = $12,225 , closer to buy / hold scenario
*in 1 year at 5x , mined total is $27,375 - $4200 hardware / power = $23,175 , better than buy / hold
*in 1 year at 10x, mined total is $54,750 - $4200 "" = $50,550 , $7450 more than just buy/hold
Check it out, let me know what I missed, (of course I know it can't be that simple, but I think the basic logic applies)
I think a better way of working it out would be to use an average daily profit to work out what would be more profitable rather than an arbitrary number of coins that way it could be used for multiple coins, so in this scenario for the mining option it would be $15 per day @ same cost. So anything over $11.50 per day profit better to mine than buy, and you will also have a mining rig left over at the end of the year to continue mining with. But as you said with the other variables to factor in its near impossible to predict coin prices or difficulty increases.