Epicenter - Learn about Crypto, Blockchain, Ethereum, Bitcoin and Distributed Technologies - Meni Rosenfeld: Mining Pool Reward Systems, Bitcoin Economics, Bitcoin in Israel

Episode Date: October 27, 2014

Meni Rosenfeld is Founder of Bitcoil and Chairman of the Israeli Bitcoin Association. Having organized several meetups and conferences in Israel, he is a very active member of the Israeli Bitcoin community. Meni has been studying mining pool reward methods for several years and in 2011, wrote a paper titled “Analysis of Bitcoin Pooled Mining Reward Systems”. This research gave him a broad understanding the reward system landscape and the issues with many of the methods used by pools. He later wrote proposals for alternative non pay-per-share methods and which would render pool hopping impossible: the geometric method and double geometric method. Episode links: Mining Pools Reward Methods Talk - Bitcoin 2013 Conference [YouTube] Analysis of Bitcoin Pooled Mining Reward Systems [paper] Summary of mining pool reward systems [paper] Analysis of hashrate-based double-spending [paper] Meni Rosenfeld's vanity thread [Bitcoin Talk Forum] Geometric method [Bitcoin Talk Forum]: New cheat-proof mining pool scoring method Double geometric method [Bitcoin Talk Forum]: Hopping-proof, low-variance reward system Mine in multiple pools to reduce variance [Bitcoin Talk Forum] Multi-PPS [Bitcoin Talk Forum] ASIC will not centralize Bitcoin mining [Meni’s blog] Blocksize Economics by Gavin Andresen [Bitcoin Foundation] Proof of Activity: Extending Bitcoin’s Proof of Work via Proof of Stake [paper] Israel Bitcoin Meetup Group Israeli Bitcoin Community Facebook Page This episode is hosted by Brian Fabian Crain and Sébastien Couture. Show notes and listening options: epicenter.tv/049

This episode of Epicenter Bitcoin is brought to you by Fairlay. Fairlay is a Bitcoin prediction market where you can place predictions on the likelihood of sporting events, the Bitcoin price, or current affairs. You earn money if your predictions are correct. Head over to Fairlay.com slash epicenter, that's F-A-R-L-A-Y.com slash epicenter to place your first bet today. Hello, welcome to Epicenter Bitcoin, the show which talks about the technologies, projects, and startups driving decentralization and the global cryptocurrency revolution. My name is Sebastian Kutjou.

And my name is Brian Falman-Krain. We're here today with Many Rosenfeld. He's the chairman of the Israeli Bitcoin Association. And he's kind of a very long-standing member of the Bitcoin community. So I'm really excited to have this episode today. I remember at my, I think, first Bitcoin conference. I went to dinner with Many Rosenfeld and Adam Back,

which was pretty great because they're having they're both too tremendously smart and knowledgeable people have big bitcoin so it was a great way to like hear all these have a lot of interesting conversations so i've been since we started this podcast i always wanted to have mani on at some point so i'm excited that's happening now and there is a lot of really interesting topics we'll get to talk about so thanks for joining us today minnie sure no problem glad to be here so we're going to talk a lot about mining today. Many is on a lot of work with mining pools and especially kind of the thinking through the fundamental theoretical aspects, the game theoretic aspects of that. We're also going to talk about Israel, his work there. But perhaps can you give to those who don't

know you some introduction about your background? Sure thing. So like you already said, my name is Mani Rosenfeld. So I live in Israel. I was I was born in Haifa, and I studied for a master's degree in mathematics in the Weitzman Institute of Science. I did work on machine learning, specifically semi-supervised learning under the supervision of Dr. Boas Nadler. And after I finished my degree, I started working as the head of algorithms research in a startup company called Similar Web, which is involved with analyzing web traffic and figuring out

which website are similar to others and offering an analytics platform to analyzing the traffic that each website gets. So I started working in similar web in August 2009 and about

a year and a half later in March 2011 I heard about Bitcoin and it didn't take a lot of time since the moment I heard about Bitcoin until I was no longer the head of research at similar web.

So, yes, once I heard about Bitcoin, so, you know, like all of us. Do you count that time in seconds or days? Well, no, it was a process, but let's put it this way. You know, so of course, when I when I first heard

the Bitcoin, and of course the first one or two weeks, like all of us, started reading about it, and a week later I created an account in the Bitcoin Talk Forum, which then was not yet called Bitcoin Talk, but it was, and I think still is, the major form for Bitcoin in the world. So I started writing there, of course, learned much more about Bitcoin in the process.

So, you know, I had, you know, even in the beginning, I had some thoughts about what I want to do with Bitcoin, and I have, I've had. several ideas but by April which was one month later I started to think about what I could do to make Bitcoin more successful in Israel and what I thought is that if people had an easy way to buy and sell bitcoins then adoption will be much greater because I were thinking mostly about businesses that would consider accepting Bitcoin but they are worried that maybe then they won't have anything to do with those Bitcons

so I wanted them to have the option to sell Bitcoins if they so choose so like a BitPay like a BitPay payment processor yeah well yeah I mean yes okay well that's actually maybe a more advanced stage because you know BitPay is really good with you know the integration

and making it seamless at the time I didn't really you know consider making it seamless. I just wanted the option to be out there, yeah, so that people will know that Bitcoin are valuable, that they are not stuck with some play money that nobody will want or them to see that, yeah, there is

an actual value for Bitcoin, they can sell it at any time, and actually the mere fact that they know they can sell it can itself make them less likely to sell it because there's no rush, right? It's They understand that Bitcoin is not like a hot potato they need to get rid of. So yeah, so by the middle of April 2011, I figured out that, yeah, okay, let's do this.

You know, I didn't know anything about starting a business. I didn't ask anyone, it just, and I also, you know, like I said, I'm a mathematician. I not really do much software development. I do it as a hobby, but not really. anything professional. So I wanted to create a website for this service.

So I learned the PhDs that I needed in order to do this and didn't ask anyone had just registered a domain, put up a site and posted on the Bitcoin Talk Forum. Yeah, okay, I'm buying

a Bitcoin in Israel. You know, come and see me. And what was this startup called? Oh, yeah. So this service was called Bitcoin,

which is like, you know, like a play on words, it's a Bitcoin, but with a TLD, C-O-I-L of Israel. So, yes,

so, you know, so ever since I started it, I had some interesting stories, but the, the point is that, that was the, you know,

my first activity in Bitcoin in Israel. And maybe I'll take, take us a few weeks earlier, to something which happened maybe one or two weeks after I heard about Bitcoin, which we'll talk about much more later, which is mining pool reward methods. Back then, yeah, I don't, let's talk about this more a bit later, but the first thing I did was develop a new reward method for mining pools,

and later on I started Bitcoin. And let's see. Yeah, and you know, so in the beginning, I didn't know who else is involved with Bitcoin. So, you know, I looked around. I didn't see anyone in Israel doing anything with the Bitcoin. Well, that's not completely accurate. There were some people who did something, but nothing very serious. So, you know, so eventually I realized that, yeah, I'm the first person in Israel.

that took Bitcoin seriously and really did some bigger projects with it. And yeah, so this continued, you know, back then, you know, I ran Bitcoin. I was involved in the community in Bitcoin token. The Meetup group, no. Yeah, and we also had a, you know, a Facebook page on, you know, about Bitcoin where we talked a bit. But in August 2011, there was this. the first Bitcoin conference ever, which was in New York. So I don't remember when exactly I first heard about it.

But I think that when I did, I thought, what does this have anything to do with me? I mean, that's in New York. I'm in Israel, so it doesn't really involve me. But by a bit a few weeks before that, maybe, I don't know, two or three weeks, I heard about it again and then I thought

this is the first Bitcoin event in the world I have to be there so you know so it's so in really fast I did all the preparations needed to travel

and they went to that conference and it was really great and you know and I thought that we need something like that in Israel as well so on our Facebook page I announced that we're going to do

a meetup and you know so I chose some location and the eight people said RSP'd but eventually only three showed up which is me and Ron Gross which was also very active in the Bitcoin community in Israel and another guy which is which is not very active in Bitcoin now so so that was the the first Bitcoin meetup in Israel three guys sitting in a cafe it's a it's a it's a story people like to tell because it grew much more since then so you know speaking about meetups specifically and afterwards we had meetups with 10 people 20 people by January 2013 there was a big

growth in Bitcoin around the world but especially in Israel so we had a meet up with 120 people and by two months later we had a meet up with 160 people in February 2014, we had the first real Bitcoin conference in Israel. It was a local conference meant just for Israelis, conducted in Hebrew. And I think we had maybe 320 people there. So that was also very nice. And on that note, yesterday we finished the first international

Bitcoin conference in Israel, which which was very nice. So there were less people, maybe 250, I think. But, you know, it was a much more major event. It was two days. It was international. It was in a much nicer venue. So, yeah, so I think...

You're talking about Inside Bitcoin's Tel Aviv. Yeah, yeah, that's true. So this conference was actually a collaboration of the Israeli Bitcoin Association and the Buzz Productions, which is the company here that produces conference. and the MacLear Media from the Instant Bitcoin brand. Of course.

So, yeah. So that was also very successful. So your meetups, they've kind of taken the turn from meetup to really small conference events, right? If you're bringing in 160 people per meet. Is that people attending the meetup? Or is that a number of subscribers on your meetup page? When you say 160 people? Yeah.

Yeah, when I say 160 people, I mean the people who were actually there, we counted them. Usually half of the people that sign up to the meetup show up. So, you know, I think there are many meetup groups where, you know, they have the number of RSVPs. And if you don't do anything about it, then after the meetup, the Meetup.com shows the number of RSVPs as the number of people attended. but that's of course not true and I make an issue of being honest and counting how many people were actually there and showing that number up

and yeah, 160 people is the number of people who were actually there, of course. Maybe we miscounted a bit, maybe five people here and there, but that's the number, that's the number. But that's really impressive. I mean, and you're doing these every month? Okay, so about that. okay, but let's, I'll just make a comment,

and there were some people that I talked with, and I told them, yeah, we had a meetup with 160 people, and that's not a meetup, that's a conference. And, you know, my rule of thumb is that in a conference, there are name tags. You know, so in a conference, it's much more, you know, maybe more formal, more big, you know, nicer. So I don't view these meetups in the conference,

I think what we had in February this year, that was a conference. You know, it had sponsors and the exhibition hall and all that. So, yeah, so, you know, but it's semantics eventually. So, okay, yeah, now about the frequency of the meetup. So in the beginning, we had the meetup every two months. Now, but lately we switch formats, you know, so we have, you know, a big conference about every half a year, but these bi-monthly meetups were not doing that. Instead, we're doing bi-weekly lectures because, you know, the meet-up that we had, for example,

in March 2013 with 160 people, it was a lot of work to do. Yeah, it was, it had lectures, it had an agenda, it has a lot of, you know, logistics and so on. So it was a lot of work to do. and we got to the point where we can't dedicate the amount of time needed to bring, to raise such a meetup with high quality standards. So we did, you know, so we did something which is a bit easier. So what we're doing now, every two weeks, we hold a meetup dedicated to one lecture on one specific topic and, you know, we have a wide variety. So sometimes it's more technical, sometimes it's more industry.

printed, so we really try to have a huge variety of these lectures. And organizing just a lecture is much easier. Yeah, you just choose a topic, post about it, and people show up, and you have the lecture, and we are used to doing this. And of course, there is work there, but it's much less than organizing a bigger meet-ups. So this is something we can have a regular schedule of once every two weeks. And so I imagine you're doing these in Tel Aviv? Yeah, indeed. So we have actually, there is in Tel Aviv.

And so people are coming from all around the country. Yes, so yeah. People are coming from all around Israel. Yeah, you know, so let's see. So you know, usually, you know, the Israel in its entirety is a pretty small country. So yes, so people from all the country can make it to Tel Aviv. But, you know, so for example,

If you have these major cities which are relatively close to Tel Aviv, such as Saifa or Jerusalem, then it's easier for people to come from there. If you go a bit further than that, then people usually will not travel all the way unless it's a really big meetup. So, yeah, and you know, and we're seeing an increased activity in other major cities. I mean, most of the Bitcoin activity in Israel is done in Tel Aviv, but we're seeing more activity in Haifa and Jerusalem. We had a meetup in the Technion, a big university in Haifa, and we're working now and organizing

a meet-up in the Hebrew University in Jerusalem. So, yeah, but it's still the case that most of the activities in Tel Aviv. Okay. And touching on the Israeli landscape, and not to stay on this topic too long because we have much more interesting things to talk about, what is sort of the regular environment like in Israel. And I think you've had some experience with this because your service Bitcoin had some problems with Israeli banks, correct?

Yeah, this is correct. Now, by the way, yeah, I agree that we shouldn't spend too much time on this topic. But I do want to say that, yeah, so eventually, you know, there is a lot of activity that we're doing in Israel. and eventually we decided where we is mostly Ron Gross and I, to make it more formal, to make it larger and more scalable. And that's when we formed the Israeli Bitcoin Association, which is a registered on profit, which has several goals.

One of them is educating people about Bitcoin, which is what we do with all the lectures and so on. By the way, every lecture that we're doing, and every meetup and so on. We're recording and uploading it to our YouTube channel. So most of it is in Hebrew, so it's not going to interest most of the listens here. But occasionally, we have a special guest in Israel.

We've had, for example, Vitalik Buterin and Peter Todd and Adam Beck, and so on. So when they are here, we have a special meetup just for them, and we record this lecture. So this is all available on the associations. YouTube channel. Now, yes, so another goal relates to what you said, which is working with the government and the authorities to generate a more Bitcoin federal regulation. So the situation right now is that there are no official policies from the authorities regarding Bitcoin,

unlike what we saw in the US, that FinCent had something to say about it and the IRS had something to say about it. In Israel, there are no official guidelines. There are some statements, for example, eight months ago, the Bank of Israel, together with some other agencies, published a kind of warning, you know, detailing all of the risks that Bitcoin are involved. So, you know, so like a caveat at Amtor and you know it created a lot of buzz and right now there's also a lot of buzz about something which is not an official sentiment by the Bank of Israel but it is a research paper conducted in cooperation with one with with a member of the research department the Bank of Israel so and this paper mostly explores the potential of Bitcoin so so yes so from actual

regulators, this is a situation that there's nothing basically. So banks do whatever they want. And what they usually want is not to allow anything to do with Bitcoin, but this is also starting to change. There's one bank which is starting to warm up to

Bitcoin, maybe even two banks now. So right now, you know, we are making do, you know, we can, someone wants to use Bitcoin, can get some, he can do it. Maybe not always is clear, but we can use Bitcoin without too much problem.

And yeah, this most covers this issue. Yeah, I mean, I think it's really, it's really exciting listening to you to see how fast and how much it has grown. And I think it's definitely very noticeable to me that there is a lot of things going on in Israel. No one meets a lot of people going the conferences from Israel today.

They're all from Bitcoin companies there. And it's interesting to hear from you how it's such a short time. It's come such a long way. And also, of course, congratulations for the work you've done there. I think that's one of the things that I find really impressive about you is how, like, you've done so much work on so many different fronts from like it, especially like educational activities as well, like from posting on Bitcoin talk or a stack exchange where, you know, extremely active.

Yeah, sure, you know, we usually enjoy doing this work. You know, in the beginning, you know, I thought about it in a business way, you know, especially regarding the work in Israel. So, you know, I thought, okay, I have a Bitcoin a business in Israel. So if I help Bitcoin in Israel grow, then my business will be more successful. Eventually, the banks back then didn't really allow me to, to continue operating and running an exchange service

wasn't exactly my foretests so that business you know it's no longer active but you know I had some momentum you know I was doing work on Bitcoin in Israel up to now so I didn't really see any reason to stop so just doing

whatever I did before and you know and about the research stuff then yeah you know I like of course I like mathematics I like doing research and so yes it's good for me that

I have the opportunity to do research on a topic that I like so yeah so let's move on a bit to some of the work you've done regarding minor reward systems which you mentioned that was kind of the first thing

you got into and I was actually we both were earlier reading through the paper that you wrote, which is quite a while ago as well, about mining reward system. It was very interesting read. I thought it was very well written and understandable as well for someone like me who I'm, while I'm an economist, I'm not a mathematician, and I'm also not particularly familiar with the intricacies of that, but it's fairly understandable.

But perhaps before we dive sort of in the technicalities and more in depth of the work you've done, Can you give a very brief introduction to mining pools, what they are, what the function is, and what the role of reward systems in that context is? Yeah, sure. So, okay, so a mining pool is basically a group of miners, mining together and sharing the rewards. I hope I don't have to explain too much what mining is. If you want, I can do that. No, I think that's okay.

Yeah, okay, so assuming there is mining, it turns out that small miners who have like a small mining equipment and they try to mine, if they mine on their own, they will have a lot of what is called variance. So, for example, if I have like, I don't know, one terahe hash per second and they look at the difficulty and so on and calculate and I find out that they will get, I don't know, a hundred dollars payday for mining, I will not actually get a hundred dollars per second. I will not actually get a hundred dollars. So, on most days, I will not get anything at all. And, okay, very rarely, I will hit the jacked pot and earn $10,000. And, you know, the way this works, which is completely random, has some unintuitive property. So, for example, people can have an intuition that if they haven't found anything in a while, then they are due to final. So in the end, it will all even out.

and it doesn't. So, you know, if you mind for a month and you didn't find anything, no one is going to give you back the work that you lost. The work really was lost. So this is a very difficult situation.

And I'd like to tell the story of how in the beginning, I myself tried to mine solo. And it was difficult because, you know, I found maybe one block every several days and it kept me up at night

I used to lie awake in bed and think did I find a block just now or no so it was on my mind and so that's why most miners prefer to have a steady payout that they will receive a payment

exactly proportional to their time mining and the hash rate and this is what mining pool introduces. It reduces that variance and so you have a regular payout based on your share. Exactly, yes. So you know, you can mathematically calculate exactly how it works out, but eventually, yeah, a mining pool can allow you to have a much more steady payment.

Now, actually, the larger a miner is the less is affected by this variance when he's solo mining. and in a pool, the variance that each miner has is as if he mined with the combined size of the pool. So, for example, if I have one teresh per second and I mine in a pool which has one pet a hash per second, then I will have the same variance, the same relative variance,

than one pet hash per second miner, which is, of course, much lower. So that's, you know, that's the beginning of the theory of mining pools, but it gets much more complicated because the pool needs to decide how to split up the rewards. And this turns out to be not a very trivial problem. So perhaps we should just point out that there are different types of ways to split that reward. I guess probably the system that was mostly used a few years ago and even now is the paper share. so you get paid a certain amount of Bitcoin

based on your percentage of the total hashing power, which is the paper share model? Yeah, yeah, you could call it that. Well, you know, I guess I should explain first a bit what a share is. So, you know, so the first step in speaking out of the rewards is measuring how much work each miner does

and this is done using shares. So a share is like a smaller block and easier block. So a small miner might never find a block, but he will find shares. And it's impossible to find shares without doing the work needed to find a block. So when the miner finds a share, he submits it to the pool, and this proves that he has done work. So this is the first step. The second step is to figure out what to do with that.

So, yeah, so in the paper, share a method, which I'm not sure if it has been the most popular, but it is an important method. So the Pools operator simply pays out the miner for every share that the miner finds. So this is in theory the best method for the minor because he has no variance at all. He has no risk. It's all very simple.

It just sends shares and gets payments. Very good for the miner. However, it's not so good for the operator because the operator needs to pay the miners whether they find blocks or not. So if the pool finds a lot of blocks, then that's great. The operator gets to keep all of the rewards. But if the pool finds less than expected blocks, then the operator will have to pay out of his own pocket.

And you can calculate the risk that the operator has, and it's a pretty big risk. So traditionally, PPS pools charged very high fees for the service. they have risk, they want to compensate for it. So the fees in the PPS pool are very high and this is a problem for the miners. So it's an alternative to PPS. People invented a method

they usually call the proportional method where, yeah, the basic idea is that we look at the time between two blocks being found. So if you look, yes, so let's put it this way. The pool has rounds, and whenever a block is found, a round ends and the new round starts. And for every round, the pay, the reward for the block

that was found at the end of the round is distributed among the people who mine do, during this round in proportion to the amount of shells that they've submitted. And this may be sounds logical at first, but the problem is that this method is completely broken because of the infamous problem of pool hopping. So a pool hopping basically means that in a pool that uses the proportional method, sometimes are better to mine than others.

There are sometimes when I can mine, and the expected amount I will get for every share is more than what it should be. So what people can do is to mine in the pool whenever it's profitable and when it's not profitable, mine solo or go to another pool. And of course, this is at the expense

of the continuous miners because there's only a given pie and if the hopers take a bigger portion of it, it, the continuous miners get less. So this is a problem that for a long time has been very poorly understood. Even when people started to understand the problem, it was considered very theoretical. No one was sure if people can do it, if they actually do it. There was this time where someone came forward and admitted, yes, I am pool hopping.

And this was, you know, like, I don't know, shocking, but it was like, why. there is actually someone who is pool hopping, but ever since then, everybody started to hop. There were easy tools that were given out that everybody could use to hop. And, you know, for a long time, people thought that they could salvage

their proportional method somehow, but eventually there is that no. It's just, it's broken from the ground up. You just need to throw it away, find a different method. And around that time, shortly before I started being involved with Bitcoin, there was this guy called Raolo, which invented a different method, which was first used by Slashes Pool, which is why I just call it Slashes Method,

which, you know, which has, you know, it was built in a specific way, which resists pull hopping. You know, I don't want to go into all the details, but the basic idea is that every, the value of every share that the minor submits decays over time. So if I,

if I submit a share now, but the block will be found a long time in the future, I will get much less for it. So this was the right direction to go in, but the method wasn't complete.

It was still somewhat hopable, and there were other issues. It was not very, you know, mathematically rigorous. So I looked at it and, you know, I thought what could be done about it. And after enough thinking, I came up with a method which is completely hopping proof. And, you know, it had specific tradeoffs,

but I believe that, you know, that hoppability is the worst. thing that could happen. And if you have to sacrifice some other performance parameters, then this is still worth it. So, you know, so in the

beginning, you know, I offered this method to slush and he said it's interested, but it didn't really have time to do anything about it. Eventually, I wrote a post about it and a few pools adopted it. And yeah, and this is

this is similar in many ways to slush this method, but it just makes a few things more accurate. And yes, so, yeah, and back then I've done some, you know, some more theoretical research about it. You know, I found out that in a way, by the way, this method was eventually named it the geometric method. So I proved that under certain assumptions,

a geometric method is the only hopping proof method, but eventually I came to realize that that these assumptions, you know, can be relaxed and you can find the methods which are better than the geometric method. So, you know, there were a few other methods. So right now, the most popular method, which is Hopingproof, is PPL and S. And, yeah, that's where, so when I started writing a paper

that simply covers all of the different reward methods and analyzes them. And that's, I think, still a useful resource for all of the classical mining pool methods. If we have some time, I will talk about the newer age of reward methods. So any questions so far? So, yeah, well, I'd like to go into, perhaps explain a bit more what the double geographic method is, how it works in comparison to the other methods that we described, and also talk about the sort of the general landscape of mining and how it's been evolving towards these sort of more just methods

where you can't cheat in just a minute. But before we do that, we'd like to take a second to talk about Fairlay. Of course, Fairlay is a Bitcoin prediction market where you can place predictions on the likelihood of sporting events, the Bitcoin price or current affairs, and if you bet correctly, you earn money. And so, Brian, can you perhaps explain how a Bitcoin prediction market works? So the main difference between a traditional prediction market and something like Phelay is that you don't, like, bet or you don't take a position against the house or the company, but it's a marketplace.

So essentially, there's always a counterparty on the other side. And failure is just sort of a middleman and to take a small fee. So that can be a lot more efficient and can be a lot cheaper. fair, especially if you have a good volume. Now, what's also interesting is to think a bit about use cases. And the today's bet, the one we wanted to point out, was on the Bitcoin difficulty. Now, I'm trying to pull this up here. Just one second.

So, for example, there's a bet that the Bitcoin difficulty is going to be higher than 75 billion, is that be for billion? Many you would know? Yeah, it's so I think so, yeah. By the end of the year. Now, it's kind of interesting

to think about why would one take a position on a prediction like that? And of course, there could be simply a gambling thing, right? You could do this for fun because you think it may be fun or not. But you can also say, instead of, for example,

buying, mining hardware, which in a sense is a bet that the difficulty is not going to rise too much, because if it rises too much, your hardware is going to lose all its value. You could say, I'm going to bet that the difficulty is not going to go too high, so you can take a position against that. Or if you already have hardware, it is a way to hedge. So there are all those different ways of thinking about predictions, which is kind of interesting, and these different use cases. you know i think a prediction market i get you know in the broader economy economics perspective predictions market are a way to for the whole market to gain information about specific

things by you know by rewarding people who generate accurate predictions so and yeah this this information that the prediction market generates can be useful when people make decisions and they can plan the economy more effectively so we wish fair day the most success. Now, specifically, you know, but in order to really be useful, the thing that you bet on should be actionable. So a bet on whether the difficulty will be higher or lower than some amount, I don't think that's a very effective way. The information you get from it doesn't really influence much decisions because it's, you know, it's an incomplete bet. You know, when people need to decide whether to invest in mining equipment,

then it's not enough for them a binary prediction, difficulty will be higher than that. I mean, what they want to know is what will be the integral of the ratio between the big and the difficulty from now to infinity. So, you know, so if you do something about that,

then, you know, it becomes a certain of a commoditized the mining contract, which is similar to what I tried to do with pure mining, which was the first deterministic mining bond that I issued on Glebs back when Glebs was a thing. So, yeah, so I think there's a lot of interesting things there. But yeah, specifically a bet about the difficulty I see. I don't see the value in that.

Well, if you'd like to bet on the difficulty level being at $75 billion by the end of the year, you can definitely do that by going to Fairleigh.com slash Epicenter. That's F-A-I-R-L-A-Y.com slash Epicenter. Sign up, deposit some money and give it to try. And we want to thank Fairleigh for their support of Epicenter Bitcoin. Absolutely. Thanks so much. Now, let's dive back into the topics. And I would suggest since we're already 40 minutes in that we don't spend too much time more to speak about the specific reminding reward schemes. but we talk a bit more about maybe the broader picture of Bitcoin mining. Yeah, okay. So I'll just briefly mention that DGM or double geometric method is one of the methods I invented. It's pretty complicated.

So far, I don't think any pool operator managed to implement it without me, you know, sitting with them and explaining exactly what they need to do. So, you know, back at the time, I thought that, D-JM was worth the effort. But right now, you know, I think we have bigger fish to fry. And also, you know, people are using people in S and, you know, a little knowledge is dangerous. So people think that they understand people in S.

So they implement it the way they think is right. And it's not exactly accurate, which I used to be a bit annoyed with at the time. But now I think we have bigger fish to fry. so the problem right now is mining pool centralization where if you look at my paper about reward methods from I think it was three years ago now so you know all of the methods that appear there have a certain aspect that you know they with these methods size does matter. The bigger a pool is, the more performance it can offer to its miners. So this causes a centralization problem. So you know, if you have one pool that is slightly bigger than the others, then all the miners will start flocking to it until it gets bigger

and bigger. And eventually there's just the one pool where everyone mines and they are unable to mine anywhere else because all the other pools are so small and offer such bad performance that it simply makes no sense to mine there. They think this is exactly what we saw with Giga-H.io that reached the majority of the network cash right a while ago. So simply because it was so big, every miner which is not completely altruistic,

simply mine there rather than anywhere else because the variance he would have mining for that pool was much lower than other pools. So this is a problem all of the classical reward methods have, which is why right now I'm focusing more on new frameworks for reward methods. So there are actually two ways to combat decentralization. So one of them, which I think I suggested two years ago, is that miners should mine in multiple pools simultaneously.

So for example, if there are 10 pools, then I as a minor split my hash rate and spend a little bit in each. And this is a win-win situation. So for the miner, he gets the same performance as if he mined in a single huge pool with a combined size of all these pools. And for the network, this is much healthier because the minor maintains the status quo. He doesn't give more power to the pool which is already in the most powerful, but he keeps the same, distribution as before so pools can compete on their merits rather than who manage to get bigger sooner so in this way small pools will still compete right it you know if there is small pool then people will still mind there not a lot but they will mind there so if so the pool can

can grow and yeah so this is i think something which is very important um which is this is something that's being done today Well, I don't know for sure. I mean, back when I suggested it, you know, a few miners said they were going to do it, but I think there's not enough awareness that this is even possible or desirable. But I think the even bigger problem with that is that pools are to commodities. So, you know, even if someone creates a software client, which the miner installs and knows how to...

to mine in several pools simultaneously, the user can't just tell this client, okay, mine for the 30 best pools and so on. So, you know, the user actually has to go to every pulse website and sign up for an account and log in and so on. Yeah, okay, I just realized that I've forgotten to plug my computer to the power, but let's hope that it will, less until the end of the session.

Anyway, so, so you know, so what, and you know, so if we wanted to be effective, we need to mine in like, I don't know, 20, 30 pools, and it's a lot of hassle to sign up to all these pools. So I think what we need to make is that pools will be more of the commodities. So you can have a client that automatically searches for pools and logs into pools and the sensor octopause without the user needing to manually intervene. So do you think this is going to be the way this problem will be solved? I mean, your solution does make sense to me, right?

It makes sense to me that this would address the mining centralization issue. Of course, you're totally right to point out. It's also an overhead, and it very much seems to be the case that miners are, you know, until something really wrong happens, nobody really sees the incentive to do something. And I think you have this sort of public situation where you think, oh, the other people are probably going to do something if it's really bad. So I will just stay put and do the easiest, most convenient thing. And this is obviously not it. So do you think the mining centralization issue will be solved?

do you think it's going to be something like that where you'd say you'd try to split up mining power among many pools or is it going to be something radically different that will address this issue? Yes, so first I want to say one more thing that actually, you know, actually, you know, the what I described just now is simple, but, you know, it's not enough.

So what I'm mostly advocating right now is something which I call multi-PPS, which is based on the same idea of mining in multiple pools, simultaneously, but it's a bit, you know, once we make it happen, it will be much more powerful, but it requires more work to do. So, like I said right now, multi-pPS is what I'm focusing on, you know, you can, currently this method exists mostly as a post on the Bitcoin Talk Forum, so anyone who's interested can look it up.

So, yeah, so I think, you know, the important thing is that we don't require miners to be very altruistic, right? So, you know, right now, when people talk about centralization, they may be suggest to use a P2 pool, which is great, but actually the performance in P2Pool isn't so good. So miners actually have to sacrifice performance in order to do the right thing of mining with P2Pool, which is why I don't think P2Pool is a solution.

Whereas in my suggestion, then yeah, maybe there's a bit more effort where the miner needs to switch to a new method but once it does, it actually gets better performance than he does right now. So we do not have to assume that minors are altruistic. And so yeah, so that's about that. So I think really this is what mining will look like. However, it should be emphasized

that this only serves the problem of mining pool centralization. If we have a lot of miners and we don't want them to create a pull which is too big, then multi-pPS is great. However, it doesn't do anything about huge mining farms which enjoy economy. Data centers that hold,

yeah. Yeah. So, yeah, so it doesn't do anything about that, but I don't think this is really a big problem because, you know, big companies have a lot of overhead. I think it's possible for at home miners to compete. You know, miners at home don't really enjoy these economies of scale, but you know, they are more efficient, they can utilize their existing

infrastructure and they can mine for other purposes than just pure profit, you know, it's interesting, it's, you know, it's cool, so, and they also are less risk of us, right? You know, a big company doesn't want to lose, but a small miner, small user can, you know, he can take this, so, so I think we, you know, so with the, you know, so the market, you know, mining from the issue exists. I don't think it's a huge problem and therefore it's not if you can solve the mining pool centralization problem.

So you think there is a future for home mining, for individual miners that just maybe some economic downsides they have in terms of having smaller scale maybe not professional

cooling facilities, etc. will be outweighed by other advantages they have as in perhaps free electricity from using, you know, using some sort of heat that's generated from something else. Yeah, there are a lot of options there. I mean, yes, so for example, especially users that live in cold countries, they might have, you know, they can salvage some of the electricity costs by using it as heat. instead of a normal hitter, you know, the effectiveness of this is not great, but it's still something that will happen.

It will help them compete. So I think we will see a combination of, you know, huge mining farms which are built in places which are cold and where electricity is cheap. I think we will have, you know, small businesses that do mining, which may enjoy some economies of scale and they're a bit more, have a bit less. overheaded in big companies. I think we will also have small miners. So all of these will be together.

So I don't think any one entity will reach a huge percentage of the network hash rate. But when I first started getting involved in Bitcoin, I guess my idea of mining, the obvious sort of scenario for the future was that every internet-connected device, every potentially, device connected to the internet would be able to do mining. This is how we would solve this issue of centralized mining is by just everything being decentralized

on every, whether it be your TV, your cell phone, your fridge, your car, what have you. Yeah, I'm not sure. Is it something else that could perhaps... Yeah, I'm not sure this will work because, you know, in order to do mining, you need something which is efficient for the purpose. You know, Bitcoin uses

Shatou 56, which is completely dominated by Essex, which I think is a good thing, but even if you use a different hash function, which is better in CPUs, still the kind of CPU that you would have in, you know, in a cell phone, I think it will be very good for this purpose. So I don't see a future where every device will mine, but I do see a future where miners will just plug in, you know, a card to their, to their computer, and it will mine, it will be built for this purpose. Let's kind of move on from the mining topic to a topic that's really related to this,

which is Bitcoin economics and the sort of economic viability of Bitcoin, when we talk about the long term, and of course that does tie in, at least to some extent, to the economics of mining. Now, I'd like to perhaps sort of segue into that by discussing briefly, a blog post, Gavin Driesen wrote last week about, I think it was called like block size economics. And in it, he made the argument that the block size should be kept limited to one megabyte. And of course, the consequence of that is that there's also only limited number of transactions that fit in there.

And in the long run, if actually more people want to do transactions and there's room in there, You know, this really drives the transaction fee price. So, you know, if there's a limited supply, this would drive up transaction fees and would drive up revenues for miners. And I guess his idea is that this would be a way to ensure miners have enough incentive to keep running their hardware. So we have a secure Bitcoin network because after all the mining reward, block reward will be dropping in the future. So, and one question that is looming, and I think is on many people's minds who think about these issues, is, will this create a problem in the long term? So what are your thoughts on this issue? Yeah, okay, so yeah, I have a few things to say about that.

Maybe we'll start with, yeah, you know, you started my mention in Gavin's post. I think the analysis there was very incomplete. I think he doesn't really consider all of the issues here. So, you know, I want to maybe start with something that, okay, maybe I'll briefly repeat the problem. If we remove all limits on the block, then, you know, the marginal cost of adding a transaction to a block will be close to zero. So miners will have an incentive to just include every transaction, but if they include every transaction, then no user would want to pay any significant fees because he knows that

transaction will be accepted anyway. So this is a tragedy of the commerce problem. So actually, it's kind of a reverse logic because in this case, if miners were a monopoly, usually we think of a monopoly as a bad thing. But in this case, a mining monopoly would be a solution to this problem because this monopoly, this monolithic mining entity could just decide what fees that he wants to do. But if you have many independent miners, then even if the best thing for the miners and for the Bitcoin network is that their fees will be low but significant, then the miner, to maximize his own profit, will accept transactions which have insignificant fees.

and this is a problem and right now it's not a problem because there is a limit on the block size but if we remove that then then it will be a problem so the solution that Gavin has suggested or mentioned in his post is

assurance contracts I don't really believe in this solution in theory it doesn't really work because whatever the situation is if the total pledges are more than are required to support the network, then no user will actually have an incentive to pledge.

He would much rather have other people pay and he won't pay. And if we don't have enough pledges, then the fact that the minor pledges will not help because it's just one part in the whole group. So I don't think Assurant contracts have a strong theory behind them. And I also don't think they have a strong empirical. support for them because you know Gavin mentioned the example of Kickstarter which

is of course a pretty successful platform but you know there so first people are not being completely altruistic I mean they do get some benefits from pledging to Kickstarter you know for the pledge itself not just the

completion of the project but they have these perks so maybe maybe they pledge to do this and also the amounts are low enough that the user can just say, what the heck,

I'll throw in $10. But, you know, so when we're talking about incentivizing miners to secure the Bitcoin network, we're talking about much bigger amounts,

and I don't think this is a place where we can trust people's altruism. So I'm not a big believer in assurance contracts as a solution. Now, as for,

you know, blocks has limit. I think it's, solves the, you know, it's the wrong solution. I mean, if, I mean, okay, I think it's very obvious that the block size limit should be more than one megabyte. The question is whether it should be 10 megabite, 100 megabytes, I don't know what, but I don't think that the block size limit is a solution to incentivizing miners because

it creates, prevents incentives. What Blocks has limited us is encouraging people to have transactions which are physically smaller and this has really nothing to do with the issue of incentivizing miners because miners just hash, the difficulty of the hashing has nothing to do with the physical data size of the transactions. So this is, you know, this is a bargaining problem between the miners and the users. Let's assume for this discussion that miners are a monolithic entity. So the miners say to the user, okay, I can include your transaction. It doesn't cost me anything.

But, you know, but you want me to do it. So I want you to pay me for doing this service for you. So, you know, there's a lot of theory about bargaining and sharply very, and so on. But the main point is that we should take whatever is the value of the transaction for the user and the fee should be about half of that. You know, I don't want to go into all the details, but the amount of fee that the user should pay should depend on how much he wants this transaction to enter the block. And again, this is nothing to do with the data size of the transaction.

So of course the problem is that we don't know what is the value of the transaction for the user and we don't want to any... You're also going to vary a lot, right, from one user to the other, even though the transaction

may look the same for the minor. Yeah, this is true. And I don't want to enter an actual verbal bargaining for each such transaction. However, I do believe that the, you know, the amount of Bitcoin sent in a transaction

is a pretty good proxy. I mean, it's, of course, not accurate at all, but I think that, you know, someone who, you know, in quantitative matters, not in, you know, emotional matters. I mean, if I sent someone a million Bitcoin, then maybe emotionally this transaction means a lot to me, but, you know, if it probably means that, you know, economically, it's not very important.

So I think someone who sends a million bitcoins will probably be willing to pay higher fees than someone who sends a millie bitcoin. So I think that the value of the transaction, the month's cent, is a good proxy to, you know, how much this transaction is worth, which is why I think a percentage fee on the value of a transaction is a good thing to have.

Of course, it has issues, and you know, you don't know how... But not a fixed percentage fee. So, okay, so. Because then we just start getting into the same type of model that we have now where you want to send large amounts of money over a bank transfer and that costs. It's very interesting that you're saying that. I've never heard someone propose that before. Of course, it makes some sense, right?

That's how all payment systems work today. I mean, I think in at least a vast majority. Yes, so I want to add a few more things. that so yeah I think that when we talk about a percentage amount then you know it also depends how much we're talking about exactly right so Sebastian mentioned that right now we pay very high fees so you know we can pay a percentage fee which is not very high I mean what I have in mind it's 0.1% or maybe even less than that so it's it would be better I should also add that you know

something which is like that but is a bit different is to if now we have a cap on the amount of data that can go into a block I think we could also have a cap on the amount of value

transferred in the block of course it will have to be some fairly large amount but what will happen if we place this cap then miners will automatically be incentivized to

charge a a percentage fee, right? So for X, for every transaction the minor wants to put into the block, then he knows that on one hand he gets some fee, and on the other hand, it costs him some space in the block.

So the equilibrium here is that there will be a percentage fee. So, you know, we could either have a hard-coded transaction fee part of the protocol, or we can have a cap on the block size. You know, it will, it's

in similar results, but, you know, the block cap size is maybe a bit more dynamic. We don't really have to set the exact fee. We will let the miners decide and the users how much fee it's worth it for them. So how do you think that would happen? Would there be some sort of consensus that this is, because it seems to me that these are dramatic changes that would be very, very difficult to get a consensus on. I mean, I think especially the one way you talk about the percentage fee or limiting the amount of Bitcoins that can be in a block. When you talk about block size, I guess that's already in discussion anyway and we'll probably have to change.

Yeah. So, yeah, so this is a major change. You know, so, you know, in general, these kind of changes to the Bitcoin protocol are pretty hard to do. because, you know, there is some status quo bias, you know, where people don't want to break a running system that is worth billions of dollars. But, you know, I think eventually, you know, I think we need to start warming up to suggestions like this because, you know, this whole thing will be relevant years from now

where the new Bitcoin created will be much lower. I mean, right now, we don't have to worry but incentivizing miners because you have this built-in inflation, which I think more than enough incentivizes this minors. So this is a problem for maybe 10 years from now. So I think we should right now start the dialogue about this and maybe eventually, you know, start thinking about making this change. You know, so for example, you know, and actually you,

if the same applies to several issues. So for example, if you look at the cost of a transaction, it has several factors. So you have the cost of storing it and propagating it. And if we want to keep this cost down, we need to put a cap on the data source of the blocks. There's also the cost for the nodes to verify the ECDS signatures, which is like a CPU cost,

which right now is not being treated at all. I mean, you can have, you know, maybe a small, physical small transaction, but which requires a lot of ICS signatures to verify. So this also has a cost, and it's not been considered right now. So I think we should add either a cap on the amount of computations needed for the block or maybe a hard-coded fee per computation. and these two things relate to the margin cost of the intersection and you have the amortized cost of transaction

which is the hashing and for that I think we need an element of either a fixed percentage fee or a cap block value so I think these are big changes and they are not discussed very much right now I think the whole discussion about transaction fees is mostly very very very naive. But I think, you know, in order to, in order for Bitcoin to work, we need to think about these things and to seriously consider big changes because I want. I think when you said that they're a little naive is somewhat normal, since we don't

really have a lot of experience with operating Bitcoin at a scale similar to the credit card transaction network. So we have very little experience in actually running this at real life size. You can model it mathematically or economically, right? And then many of these things, I guess that's a positive thing, or a Bitcoin are quite predictable. You know, I think both theory and practice have the role. I mean, we will be much wiser when we have much more empirical experience.

However, we shouldn't leave out the theory, and we shouldn't wait to have this empirical experience. I think we should research these problems and I think it's very good that there's a lot of work in academia being done on Bitcoin, you know, a lot of the research being done is

very good and it highlights very real issues. So I think we should, you know, we should do both. I mean, you shouldn't just wait, we should actually tackle this problems and first understand the problem, then offer all the possible solutions, then

choose the best solution and I really hope that whatever the best solution is it will be eventually make its way into Bitcoin because I would much rather have Bitcoin succeed than some other alt I think that the people who invested in Bitcoin early on should you know gain a return on their investment and all their contribution to the success of cryptocurrencies and think you know ultimately the power to change bitcoin is in the hands of the users and they hope they will be smart enough to agree to changes that they can help bitcoin you know scale up and they succeed in the future i think that's a great uh word to end up exactly yeah well many thanks so much for joining us today

and thanks for you know diving into these issues now we've talked about quite a few things from israeli bitcoin association to the meetup to etc um we will link to those in the show notes so if you want to read about, or if you want to read his paper on mining pool reward systems and those things, we will link to those. And you can find those on our website or on the SoundCloud page in the posting. I just want to say, I also really enjoyed talking about these things. And we don't really get to talk about the really go in depth and technical topics very often, I find, Brian, I know if you'll agree, especially lately.

and it was interesting to do the research for this and I guess perhaps realize that a lot of these topics, I'm not very fluent in them. So personally, I thought it was very enriching to do the research and then to have you on to discuss them. Sure, thank you. Absolutely. Yeah. And I think you're very right in pointing out that these are issues that they're sort of far away because we have this block reward. because at the moment still Bitcoin is barely used economically, right? A few people are actually using it to pay for things.

So these are all sort of issues that are on the horizon, but it's really important that we think about those today. Yeah. And the challenge will certainly be. I think you also very right in pointing that out that it is possible in my eyes that we will see the right solution, but that it will not be possible consensus to move to the right solution.

So that's definitely a danger in Bitcoin. Hopefully that won't happen, but it's something a danger we have to be aware of. So, yeah, well, thanks so much for joining us. And now also we have another hangout coming up next week. And that is on 8, like 1,800 UTCs, 6 p.m. UTC, I don't know what that is in time zones that I actually used. Yeah, okay, sure.

So yeah, so it was great talking with you guys. Thanks for inviting me. Yeah, thank you so much, many. But yeah, so the hangout next Tuesday, no, on Wednesday, October 29, we have David Johnson on from a big angel and master coin. So we'll talk to him about his activities. and thanks to you, our listeners, thanks so much for listening. You can follow us on Twitter at Epicenter BTC.

You can also support us with your donations, which we very much appreciate. At epistentabitcoin.com slash tips. And you can also leave us in iTunes review, which we appreciate helps new people find the show. So thanks so much and we'll be back next week.