Tuesday, April 1, 2014

Proof of Stake (PoS) - Examining NXT and MINT

"Proof of Stake" mining, or "Piece of [Bleep]" mining? This is probably a subject to tackle in more depth, but the key tenets of PoS mining are that you secure the network (those with a "stake" are less likely to compromise the network security and fork the block chain) and you don't have to "waste" lots of power. Many coins have tried this (NXT, MINT, THOR, ZEIT, BC... probably a bunch of others as well), and really the only major success story I can think of -- and I'm using that term loosely -- is NXT. Let's talk about a couple of them, specifically NXT and MINT.

The initial IPO to distribute all NXT coins was around 21 BTC, so at the highest price of BTC that would be roughly $25,000. Now there are 1 billion NXT (initially distributed among the IPO holders) worth approximately 61090 BTC. Seriously? Yes. And that's why every new coin these days seems to be trying an IPO. "Golly -- look at how well NXT did! From the IPO price to the current price is an increase of 2900+!" Who wouldn't like turning $1 into $2900 overnight -- or $1000 into $2.9 million? That's the "success" side of the story, but now that NXT exists, what are people doing with it?

I don't know how many of the initial NXT coins have been given away -- I received 3 NXT from a faucet at one point, which is basically worthless -- and frankly the whole Java-based "forging" client thing is terrible. With the source code now released, hopefully we'll see some better clients. Then again, why build a better client if something isn't really worthwhile? That's the difficulty I'm having. So let's get back to Proof of Stake mining/forging/minting/whatever.

With NXT, your chance of "forging" a block is based on how many coins you have that are "active" (1440+ confirmations since they were last transferred between NXT wallets), along with how many wallets/coins are active on the NXT network. Let me pause for a second to say that I admire the tongue-in-cheek quality of calling the creation of money "forging"; I'm not sure if it's supposed to be humorous or serious, but I choose to view it as the former. Anyway, let's assume that only about 10% of the NXT coins are active as an example. That means 100,000,000 NXT are actively "forging" and your chance to forge a block is the number of NXT you have divided by the active number of coins, times 1440 (the number of blocks per day).

You can find out how much of the NXT network is active by looking at the "target" icon next to new blocks in the NXT client, which lately is ranging from as low as 400% to as high as 7000% for the blocks I'm seeing. Divide 100% by that number -- so we're looking at 0.25 to 0.0143 -- and we get a range of 1.43% to 25% of NXT coins actively forging. That means somewhere between 14,300,000 NXT and 250,000,000 NXT are online right now. Let's just go with 10% being active, though. If you happen to be holding one BTC's worth of NXT (around 16000 NXT), then you should forge roughly one NXT block every four days (using the 10% active estimate).

So what would that actually earn? Well, that's where NXT gets a bit difficult: you only get paid the transaction fees for the block you forge, and quite a few (most even!) blocks are empty. Basically, you could earn as many as 255 NXT for a block (if all 255 transaction slots were filled), or as little as 0 NXT for a block (far more common). If 10% of blocks have one transaction (which is at least a reasonable estimate), and if you have 16K NXT and forge a block every four days on average, that means in one year you would forge 91 blocks and earn... 9 NXT. That's an interest rate of approximately 0.05% -- about what most rip-off bank savings accounts give, which is hardly worth the effort involved. Having more NXT doesn't help either -- if you have 160K NXT and forge 910 blocks in a year, you'd earn around 90 NXT, which is still only 0.5%. Yuck!

Perhaps NXT is a bad example, though, as you only get paid transaction fees, and those are quite small (1 NXT). Let's take a more popular coin: MINT. The Proof of Stake payout for MINT is 20% the first year, 15% the second year, 10% the third year, and then 5% for the fourth year and beyond. Currently, after the initial PoW (Proof of Work) scrypt mining phase, there are 18.737 billion MINT in existence. Hold MINT in your wallet for at least 20 days and they become eligible for generating PoS blocks. It's not quite clear how frequently you'll generate a PoS block on MINT (at least to me), but for the first year you are paid on the "coin age" with a target of a 20% annual increase. A 20% savings account would be awesome, right? So what's the catch?

If you hold 1 BTC worth of MINT right now (5.9 million MINT), in one year with no additional mining you should have 7.1 MINT. Neat! Except, the price of MINT is questionable to me even at 16 satoshi; with 70 billion coins planned, why shouldn't we see eventual prices of less than 1 satoshi? THOR only has a target of 15 billion and it's at less than 1 satoshi (though still in the mining phase), so it's not too crazy to think this way. ZEIT is likewise trading in the 1-2 satoshi range (and ZEIT is basically just a total rip-off of MINT, if you're wondering). But even if MINT only drops to 10 satoshi, it would take more than four years of Proof of Stake mining to make up for that drop.

So there's the question: why is Proof of Stake better than Proof of Work? It basically incentivizes people to hoard coins, and if no one uses coins in trading, buying, etc. the only people holding coins are those that already invested in the cryptocurrency. What that usually leads to is a bunch of early adopters spouting off about the "one true coin" and how everyone should by ZEIT, MINT, NXT, whatever and then hold it so they too can starting minting/forging/foraging/whatever. Those who actually benefit are unfortunately the usual suspects: the developers that take a 1% (or 2% or even 3%) share of the total number of coins ever and then dump those at the first opportunity and disappear with a large pile of profits.

TL;DR (that's "Too Long; Didn't Read): I don't get why Proof of Stake minting/forging/whatever are actually going to be worthwhile long-term. The early miners, developers, and IPO share holders will dump at the first chance they get (assuming they can first successfully pump the coin, of course), while the hoarders wait for PoS to kick in. But by the time PoS is really doing much the value of any newly created coins will be in the toilet.

Is anyone out there not holding a PoS coin that actually thinks this is a better solution than PoW coins? Part of the real draw of Bitcoin I think is the massive amount of hashing power it requires I think -- it's using the most power, and it has the highest value. What am I missing here that makes PoS coins "special"?

The only bright point is that minting/forging doesn't require a lot of electricity, but even "low power" coins like MINT and ZEIT still use CPU processor power and require your system to be on. Oh, sure, transparent forging would allow you to only power up on occasion, but that's just more work. "Oh crap, I didn't power on my system and missed my block, and now I got penalized! I guess I'll just run the system and software 24/7." On a typical desktop, I'm seeing a pretty consistent 5% CPU use out of the MINT and ZEIT wallets. That means if my system uses 35W and I leave it running 24/7 "just to be safe", rather than turning it off half the time, I'd be using 175 kWh per year extra.

That's only $17.50 or so for a year of minting, but the potential earnings from MINT are only $0.0000768 per MINT right now (and dropping). That means to at least break even on power and earn $17.50 in a year (at 20% interest and 16 satoshi), I'd only need about 1 million MINT (not much really), and in four years when we're at 5% interest I'd still only need 4 million MINT. But my bet is that we'll see coins like MINT continue a downward spiral, because no one is really going to support 20% interest rates per year, or even 5%. If the price drops to 1 satoshi (and actually stays there), you have to hold 18 million MINT to break even.

Can you think of better things to do with 3 BTC? I sure can... like just sit on it and wait for the next time the price hits $1000 and you've doubled your money. That seems a far safer bet than investing in MINT or some other PoS coin and hoping for 20% annually. Or put another way: I wouldn't bother with minting/forging unless you happen to already have a significant number of coins.