Each 2016 blocks (which is exactly 14 days if the block spacing is exactly 10 minutes), the difficulty of the useless mathematical tasks that the miners have to solve is reevaluated to keep the frequency at "one block per 10 minutes". The nearest readjustment will be tomorrow, the block 312*2016 = 628,992. The change to the difficulty will be just some modest 1-2%.
However, each 210,000 blocks which is about 3.993 years – it's a nice numerology appreciated by Satoshi Nakamoto that 4 years have 2,104,000 minutes, close to 2.1 million, and that's where the limit "21 million Bitcoin" comes from, after you multiply 210,000 by the geometric series 50+25+12.5+... = 100 – something more dramatic is happening: the halvening. The "lump sum" part of the reward for one block drops by 50%. So the Bitcoin started with a BTC 50 per block, got reduced to BTC 25 in mid 2012, then to BTC 12.5 in mid 2016, and will get reduced to BTC 6.25 on next Monday or Tuesday, see the halvening countdown.
For each block, the miners get those BTC 12 (or BTC 6.25 from the next week) plus the fees for each transaction. If they want the transactions to be confirmed soon enough, the senders of the Bitcoin need to choose a competitive "fee per byte of transaction" so that the miners choose them and maximize their profit (the block has a fixed maximum size, not much above 1 MB). Right now, the total reward for a block is some BTC 13 in average. It means that only BTC 0.5 is paid for all transactions in a block (if the block is full, there are some 3,500 of them in average; 3,500 transactions per 10 minutes means some 5-6 transactions per second, that's the Bitcoin limit).
The miners are only making a small profit if any right now and things will get discontinuously harder. Imagine that the Bitcoin remains around $8,800, the current price, for a week. Right now, a miner gets $8,800*12.5 = $110,000 per block. This will suddenly be reduced to $55,000 per block, not enough to pay for the electricity for miners.
So two things will happen: First, some miners will stop mining if the fee part of the reward (now some BTC 0.5) remains negligible. This will mean that the frequency of blocks will fall dramatically. I estimate the drop of that frequency to be 80% (plus minus dozens of percent) i.e. by a factor of five. (The drop may be less dramatic because some miners seem to be religious believers who will do this stuff even if they lose the money; and some miners are kids using the electricity paid by their parents.) Second, because the queue with the unresolved transactions will pile up, people will have to increase the fees in order for their transaction to be confirmed soon enough.
I estimate the average fee to increase from some $2 now (24,640 Satoshis times $8,800, one Satoshi is ten nanoBitcoin) to $20 or so which will happen in a few days after May 12th. How did I get the estimate? Well, one block takes the same amount of electricity and the miners just need to get above $100,000 to make a profit. Assuming the $8,800 Bitcoin price, $100,000 will still be BTC 11. But they will only get BTC 6.25 from the fixed amount. They need to get the remaining 5 from the fees, so the price for the fees will go from BTC 0.5 to BTC 5 or so (per block).
At some moment, the queue will be long enough and people will become rightfully hysterical that they won't ever be able to sell their Bitcoin again. So the sale will start. The drop of the Bitcoin price in dollars will reduce the lump sum (BTC 6.25) expressed in dollars which will increase the need to increase the fees further. A vicious circle.
OK, important blocks are
628,992 = 312*2016
630,000 = 3*210,000
631,008 = 313*2016
Note that both gaps are exactly 1008 blocks i.e. 1 week if the block were always 10 minutes. So the dates of the three blocks above would be May 5, May 12, May 19. However, when the reward drops next Tuesday, so will the mining activity. So for a while, it will take much longer than 10 minutes to mine one block. The week planned as May 12-19 may get significantly extended.
It would swell to 5 weeks if the mining activity dropped by 80%, the number I used above.
In reality, the 1-week period will be inflated less because the fees will grow, and once they grow close to those $20 per transaction, miners will be OK enough with mining because they get their revenue from the transaction fee which will replace the erased one-half of the lump sum.
The average of 1 and 5 is 3. I estimate the week May 12-19 to be inflated to 3 weeks, i.e. up to June 2nd or so. At that moment, the difficulty of the mining will drop a little bit (or a lot) in the readjustment. What can happen in those 3 weeks or so? Will the hysteria be immense? I am not sure. There is a potential for a non-event; and a potential for a much worse evolution than I described. Hypothetically, the mining may stop completely and the week will be inflated indefinitely, and the block 631,008 will never arrive again.
But independently of these mining difficulty-and-reward games, there is a simple reason why I think that the Bitcoin price will grow before the May 12th halvening; and drop right afterwards. Why? A huge fraction (if not an overwhelming majority) of the Bitcoin "investors" are obedient brain-dead sheep whose independent thinking is worse than that of your puppy (yes, carnivores and predators are generally brighter than their lunch, i.e. herbivores and vegetarians). And they're getting the e-mails similar to one that I just received from Bittrex now:
You have a great opportunity to buy the Bitcoin before the halvening.As I discussed in a similar article in January, the underlying meme is that the Bitcoin price in dollars will double because of the halvening, therefore keeping the lump sum reward for miners constant in dollars. It would make the Bitcoin system survive the halvening smoothly, so therefore it will happen. ;-) But that's just a wishful thinking and changes of the actual price are dictated by mismatches in the supply and demand, i.e. by the laws of economics, not by the superstitions of the Bitcoin religious cult.
And these e-mails manipulating the brain-dead sheep make it rather clear that lots of them will buy the Bitcoin right before May 12th (the hype!) and they will be pre-scheduled to sell soon afterwards (because they expect the Bitcoin to go to $20,000 again). So there will be an excess of buyers before May 12th and an excess of sellers after May 12th. This may drive the price up in the week before May 12th; and quickly destroy the Bitcoin value after May 12th, regardless of the question whether the transaction queue will be getting substantially long. However, if the Bitcoin price goes down, the growth of the transaction fees over the value of the Bitcoin will go up even faster, increasing the fraction of the Bitcoin wallets that are worthless because they can't even pay for the fee.
On December 9th, 2017, I told the TRF readers to sell the cryptostuff (largely because the futures started to be traded and the shortsellers could create new "supply" of the Bitcoin for the first time). One week later, the Bitcoin reached its all time high nor far from $20,000. It's still true 2.5 years older. So for the Bitcoin holders, my recommendation was precious.
While I have been skeptical about all aspects of the Bitcoin and similar cryptocurrencies in dozens of posts, this is only the second blog post that offers an idea (with no guarantees etc., I hope you understand that) to short the Bitcoin at a particular moment. If that recommendation turns out to be loss-making, my track record will change from 100% precious to 50-50. ;-)