Flaws:
1. The equipment doesn't exist, no one has ever seen evidence that it is real, and there is a very large possibility you will get nothing for your money.  What is the ROI for that?
2. If you do get it, it will be many months from now.  Difficulty will definitely be vastly higher so your earnings will be a fraction of what you estimate.
3. Price can go down in the face of rising difficulty, it has happened many times before.  Look at long term difficulty charts.  Difficulty ALWAYS go up.  Price can and does go down for years at a time.

Perhaps you should consider that mining has all of the risk of buying coins, plus the unique risks of mining on top.  If the coin price goes down your mining margin can go negative and you will never recover your investment, plus you will lose your labor and operating costs.  That happens while a coin investor still has most of their original investment available to them.  Even worse, difficulty growth can mean that even increasing price doesn't help you.  Bitcoin could triple while your money is waiting for equipment to be shipped to you.  You may make a $ profit, but it will be pennies compared to the gain a simple investor makes.

If you haven't comprehended this already you definitely need to rethink your investment strategy.  Or at least learn more before burning your money.

buy 50k of coins and stop wasting your time.

you will spend 80% of your funds locating this 'ideal' location and setting up a facility.  Then you need to mine 5x the coin costs from your miners after expenses to match what you would get from just buying coins now.

Most mining equipment will not earn 1.5x the sale price in coins over their life cycle now.

Not at all.  If you are mining at a loss, that means you could buy more bitcoin on the open market than you earned. Plus your tax position is much worse and you have to spend a significant amount of time managing your gear.

Note that the time period I discussed covered 3 boom cycles and an extended bust.

It really isn't very good advice.

I wish I had just bought bitcoin and held it with the money I invested in Bitcoin mining gear back in 2011.  I'd be shopping for a used yacht from a distressed Saudi prince now.  Plenty of people even then were arguing that mining was a negative ROI in bitcoin terms even back then. (Vladimir stands out for this particularly in my mind).

Bitcoin has been kind to me, but I wasted a lot of money, and time pursuing mining when it was vastly more profitable to just buy coins.

Not at all.  The transaction I was talking about had several hundred inputs.

And paying fees that could result in delays of days for confirmation isn't my idea of a solution.

How do you justify this story?

SegWit was the corporate takeover.  It was pushed by Core developers working for Silicon Valley startups that want to push the transaction fees into their payment channels.  It definitely favors corporate interests.

The 2x upgrade will increase the blocksize, which should lower transaction fees, and improve confirmation times.

Tell me exactly how a larger block size will favor a corporate takeover of Bitcoin.  Keep in mind increasing the block size was promised by the lead developer of bitcoin EVERY TIME scaling issues were raised.

We were promised a fix to the scaling problems.  Segwit hasn't fixed anything.  Increasing the block size will fix the problem. 

As far as I can see, not implementing 2 Mb blocks will favor the corporate takeover of bitcoin, assuming bitcoin doesn't just fail entirely over the issue.

I doubt that.  A single input transaction is costing between 1 and 2 dollars recently. 

The transaction I referenced had fees of 4 satoshi per byte.  Any lower and it won't even get into the mempool.

Kano summarized it quite nicely.  The size of your transaction depends on the number of inputs (and outputs), and fees are charged by size.  When you have an address with a lot of mining input, a transaction to consolidate them is very large, and very expensive.

As Kano's example suggests, you need to have hundreds of Th online to make your payments from the pool worthwhile.  Otherwise you lose too much in friction to transaction fees.

Now that the panic has subsided; Kano can you get your payments system working?

Since the Core fanatics seem determined to abort the 2x block size increase into another stupid fork, we will continue to not have a scaling solution.  These tiny inputs are killing me, and we are a pretty huge miner on your pool.  I can't see how most folks can even spend what they are getting from your per block payments.

I sent a transaction from a mining address last week and even manually setting the fee to a bare minimum, it cost $170 to process the transaction.  If I had used the recommended fee it would have cost over 1 BTC! 

I know who they are and they have no experience in the field.

So far they seem to be pumping their stock price via press release.

You need to tune your nose for scams if you are going to survive in this world.

Sure, because using a scammy ICO process adds enormous credibility to a company with no track record of delivering who is making highly questionable claims to begin with.

Generally speaking power efficiency improves by 15-20% per process generation.  Do you really believe these guys will get a 300% improvement from 2 process nodes versus a company that has done 5 generations of design optimizations already?

I would love to see competition in the hashing market.  If it comes it will be from an ARM or NViidia, or AMD, or Intel.  Not some unknown player who immediately starts pumping their story and heading to the scammiest possible way of raising money.

Yes it is, and it exposes you to paying taxes on earnings that no longer exist if bitcoin goes down. 

Judging by the recent lawsuits against Coinbase, it seems that the IRS is very interested in auditing bitcoin miners.  I would get your numbers in order...

Quint can you cite this ruling?

The last position I saw the IRS taking was that you were required to recognize mined bitcoin as income when it was generated and then take a capital gain / loss when selling.

The last time I looked there was a pair of river run plants for sale in North Carolina.  I think the total generating capacity was under 1 MW total, and it sounded like actual generation would be seasonal.

In practice, I expect that running the generating plant would be a full time job.

It's a nice level of rigor to hold in a pristine environment.  However, in the presence of enemy action there are no coincidences.  You wouldn't last long as an officer.

Consider that at current prices $3.6 billion dollar of bitcoin will be mined. 1% of that is $36 million.  Do you think people would fight tooth and nail over 36 million dollars?  It simply cannot be assumed that you are not exposed to people with malicious intent.  By the time you can confidently say using statistics that something is wrong the game will be over.  This is the real reason no sane pool operator runs PPS.

That said, I expect Kano is just having a run of shitty luck.  In his shoes I would be considering what an enemy might be doing, and how to detect it just in case.

It has been nearly 4 years since BTC Guild and Bitminter were hit by block withholding attacks.

In both cases the pool operator ignored the situation until I started screaming.  To Eulethria's credit, he at least following up on the obvious test I suggested of checking for large users who had never solved a block.  Dr.Haribo never bothered to respond except for an ill advised attempt to attack me, which he quickly withdrew when he realized I had been the largest hasher on his pool for 3 years.

I don't understand why pool operators haven't built in defenses for this kind of issue.  Stratum could have been revised to allow for validation work to be sent at random intervals.  There are a few other options I can think of as well but I won't speculate.

I doubt Kano's current run of bad luck is the result of a withholder, and it very well could be just bad luck.  If it was me, I would be creating a new payment address to hash against.

Bitcoin proof of work is a $2B / year industry.  If there are vulnerabilities someone will exploit them.  It would be a worthwhile exercise to make a list of attack scenarios and then consider defenses / tests to eliminate the possibly that each exploit is in play.

If you can find somebody that will buy them, I've got 150 more for him.

I don't bullshit, and you ought to look at my history before you try talking shit.

50k is either too much or far too little depending on your viewpoint.  It's too much to spend on a hobby project, and it's far too little for commercial scale.

Lets say OP is foolish enough to take your advice and chase the finished Scyrpt mining gold rush.  Lets make a budget.

L3+  $2280 (oops looks like you're paying all the profits to Jihan in advance)
Power Supply $100
Shipping    $80

For $2460 a box.  Great!  I can get 20 for my $50k.  Except 20 takes 16 kW which is way more than you are going to run or cool in any ad hoc location.

So now you need to find a location to rent or buy and modify the electric and cooling to suit.

Whoops.  Now I only have about $25k to spend...  And those 10 boxes aren't even going to earn enough to cover the rent and electric bill, let alone pay back the $50k you are out of pocket.

What open source pool / proxy software is available for Scrypt?