99% chance that its not going to be that profitable until Mid 2018, maybe till the end of the year however it might.
Let me put it to you this way
At the end of 2013, a Radeon R9 280X made about $10/day in Nov-Dec 2013. $7.50/day in Jan 2014. $5/day in Feb 2014. $2.5/day in March 2014, $1.50 in April 2014, in the summer of 2014, it made ~$1/day, at the end of 2014, it made $0.25/day.
Why? Because you are not the only one "thinking" of this idea and got money to throw around. You got competition.
Where exactly are you getting these numbers from? An R9 280x mining Ethereum TODAY on $0.10/kwh power makes $2 per day.
2013-2014 would have been the GPU days of Litecoin mining, then when Litecoin ASICs started showing up move to X11 (DASH) for a short while.
Difficulty was a TON lower at the time - my HD 7750s were pulling in over $1 NET a day each for part of that timeframe, the R9 280x would have been pulling in ballpark 6 TIMES as much (4x cores and quite a bit higher core clock).
Trading is a crapshoot. Way too many periods there was NO profit to be made, only varying amounts of loss unless you got LUCKY on timing a small spike.
Mining is a lot less of a crapshoot, if you don't go overboard at the wrong time and keep your overhead low.
There are over 2 Million GPUs mining ETH - not all ETH mining GPUs manage 30 Mh/s after all.
ZEC has ballpark 1 million, might be somewhat less depending on how much of ZEC mining is high-end NVidia driven (GTX 1080/1080ti are GOOD on ZEC, so-so to poor on ETH).
THEN you get to the small fry that might manage a half million between them all combined.
ETH going POS will hammer profitability - but not 100x and it won't be all-at-once, given the current plan to "phase in" PoS in multiple steps.
The "Ice Age" driven longer time between blocks is also going to start seriously impacting profitability sometime in the next month, it's ALREADY had a noticeable impact though not a big one.
That will ALSO tend to make the "get out of ETH" transition a more gradual thing.
Getting into mining NOW, due to insanely inflated GPU pricing and the serious profitability DROP on almost all coins the last week and some, is a major risk - one of the hedges of mining is that you can get MOST of the value of your rig back (as long as it doesn't die) by selling it when profitability dies - though as many rigs as are involved this time around, that might take quite a while to get a SIGNIFICANT percentage back - and the inflated sale prices on GPUs currently make it a LOT tougher as the GPUs are the large majority of the cost of most rigs.
It's much less of a risk for those with VERY VERY LOW electric cost, as we'll retain profitability longer than anyone else - but there will come a time that 3c/KWH or less electric farms will be about the only folks showing a profit - and right now it's looking seriously like a race between that point and the "final" move to full PoS as to which is going to kill ETH mining first.
On the other hand, this time around there is at least one OTHER coin with a fairly high market cap and a high network hashrate as part of the "profitable coins to mine" basket, and quite a few smaller ones, so it's not going to be nearly as much of a bloodbath as the end of Litecoin GPU days - and given ETH plans on POS I don't see any real probability of anyone building an ASIC for it to accelerate the collapse.
ZEC on the other hand.....