You simply cannot assume risk away. Don't give me that 7 days stuff. BTC prices are incredibly volatile, and they can be $20 or $200 an HOUR from now, let alone 7 days from now. How much money would you make (or lose!) if you plug in $20 instead of $140 for BTC prices over 7 days at constant difficulty? If you plug in $200? I feel like this is the third time I've had to make this point and you seem unable to comprehend it.
Actually no. I understand the risk perfectly which is why I have been recommending people to buy cards and you have not for almost a year now. People who listened to me made $ and people who listened to you didn't. I'll break it down for you again.
The risk occurs ONLY until the point when you break even on the resale of the used hardware considered against generated profits and resale value of game coupons. After that, there is no risk because if you were to sell the videocard after that point, your loss is $0. If there is no financial loss, there is no risk. The key question then is how many days will it take for HD7970 OC to make enough $ to offset its short-term depreciation should bitcoin mining fail and the user does not need the 7970 for games?
There are a couple ways to think about risk. First, we do not know what the price for bitcoins will be the next 7 days. One approach is to assume that the price is fixed at or around today's level for 7 days. It doesn't matter at all what the value of bitcoins is after 7 days because we will break even in just 7 days at today's prices (in this case it means the coins generated + resale of game coupons would completely offset the loss of value due to depreciation of the hardware) if the price of coins remains at today's levels. Beyond 7 days, if the price remains at today's levels or similar, we start making $$. After that point, it doesn't matter at all since if bitcoin price crashes, you sell the card and lose no $.
The second approach considers that we cannot predict the price of bitcoins at all. In this case we assume the worst case scenario which means 100% probability of not making any $ on bitcoin mining and thus taking on the loss of selling the used/opened hardware. Ie., We assume that bitcoins crash immediately at hardware installation time. Then the total risk is not measured in break even days but in actual monetary terms associated with the loss of value on hardware resale, net of positive cash flow from the resale of the game coupons. It would then be equal to the depreciated value of the 7970 (say $50 since the card isn't new) less the value of the 2 games, at least $20. Your total risk is measured as a financial loss of $50 - $20 = $30. This is your worst case scenario.
No matter how you look at it from a time perspective or financially, the risk of purchasing a 7970 for bitcoin mining is very low, the lowest it has
ever been. You do not even need to know the price of bitcoins over the next 7 days because if bitcoins crash to $0 in 1 min from now, you sell the newly purchased videocard and the game coupon and your loss for trying bitcoins is only $30.
Since I already quantified the worst case scenario, let's look at some other possibilities.
Let's assume a 90% probability that bitcoin mining fails and 10% that the price is $100 for a month. A single 7970 @ 1150mhz makes $147 after 5% pool fees, or $100 after electricity costs (assuming 15 cents per kWh).
Expected Return per 1 card = 90% * ($30 loss per card) + 10% * $100 gain a month = -$17
I assumed a 90% chance the bitcoin will crash, a 400W of total electricity consumption and the total expected loss is still only about $17.
What about a more realistic 50%-50% chance since we do not now the direction of bitcoins?
Expected Return per 1 card = 50% * ($30 loss per card) + 50% * $100 gain a month = $30
So please do not tell me I didn't quantify the risks. The math supports the view that a user should keep buying 7970s (assuming no other component costs) because the risk of loss on the card itself is very little ($30) against the upside potential of it generating profits. If a user cannot afford a loss of $30 or does not know how to sell videocards, then bitcoin is not for them in the first place.