A Comparison of Inflation and Reward Pools in Steem, EOS, Everipedia, and Karma

A Comparison of Inflation and Reward Pools in Steem, EOS, Everipedia, and Karma

Written by Ben Sigman

Who benefits? Who pays?

The “free” economy of technological services comes at a cost. As we have learned from Facebook and Google — our behavioral patterns online can be tracked and used to create anonymous profiles to steer advertising and messaging. When you the video about 80’s cartoons, have a lot of friends in their 30’s who are into animation — you get ads on your digital news site about animation products… In some cases, you talk about yogurt, and it seems like the apps are listening because next time you open Facebook, you see a yogurt ad. Famously, if you are not paying for the product, you are the product. If you’re not paying, then you are being leveraged to get someone else to pay.

For most of us, this may not matter much because our “data” is not incredibly valuable. But for people who are able to attract and maintain large audiences of followers online, this has become a very attractive career. “Influencers” have become micropreneuers, developing businesses in many types of niche community-based economies— health & wellness, parenting, adventures & travel, games, food, toys…

Some of these influencers have become incredibly successful, but together they have made the media distribution giants — Google (YouTube), and Instagram (Facebook) — into some of the most powerful companies in the world. These tech giants have control over — not just what people see and hear — but also, the success or failure of the influencers who power their platforms with content. This goes beyond their algorithms... There are many cases where content creators’ ability to earn has been shut down by the centralized organizations for unknown reasons. As you can see below, over the past 2 years on the trend in Google search for the word “demonitized” has slowly crept up in popularity as more and more creators fall victim to this censorship.


Rise of Decentralized Protocols

This is part of the reason that the decentralized systems which offer an alternative… Ad-free, reward-based, self-funded cryptocurrency ecosystems have gained a lot of traction and attention over the past few years.

I’ve been doing some research on how a few blockchain-based protocols and dApps distribute value to their networks. Each of these protocols allows its users to earn tokens based on some activity which provides value to the network. Specifically we will be looking at Steem, EOS, Everipedia, and Karma.

Each of these “tokenomic” models have an inflationary base-currency which pays out inflation to a pool of rewards. 3/4 of the examples here reward based on content voting (EOS does not have any content voting mechanism for rewards, just voting for Block Producers). If content is deemed more valuable by people’s votes of attention, then the creator of that content is incentivized to produce more content. Even the people who vote for the content get rewarded. What better way to incentivize people to separate the signal from the noise?


First, we will look at the Steem model from a very high level. This is Dan Larimer’s second project (before EOS), so it’s one that I am particularly interested in since I have been very involved in the EOS ecosystem for the past few months.

Steem, and especially Steemit is a brilliant idea. Steemit is a decentralized media platform. It’s provides a way for authors and curators to monetize their content and also make decisions about the governance of the protocol. I have experienced Steemit to be an incredibly positive environment (which is rare on the public web). Crystal Rose and I agreed yesterday that Steemit is, in our experience, the least trolled platform on the web.

Steem Inflation Model (Steem Bluepaper)

Screen Shot 2019-02-22 at 11.28.53 am.png

I think of this as a “deflationary inflation rate” since the rate of inflation goes down. Steem uses a constantly decreasing rate of inflation which is not tied to the amount of activity on the platform. The benefit here is that it creates a deflationary pressure on the token inflation. At the current price of Steem($0.286752/1 STEEM), there is about $20,000 of new Steem tokens minted and distributed daily.

Steem Reward Model (Steem Bluepaper)

Screen Shot 2019-02-22 at 11.42.44 am.png

In Steem, 75% of the daily inflation rewards go to the creators and curators of content, then a smaller amount (15%) goes to the vested (staked) token holders, and (10%) goes to the validators of the chain called Witnesses (think EOS Block Producer) or, as they are known on PoW chains — MINERS. These Witnesses are elected by the holders of the token through elections.

Criticism of Steem

However, in it’s nearly 3 years of existence Steem has also had it’s issues:

Governance — vote-buying and monopolization by whales.

Content — malignant content (kiddie porn). There is even a site that lets you find bots to create fake value on content https://steembottracker.com/Source: The Cold, Hard Truth About What Happened to Steemit.

Example of an ad from  https://steembottracker.com  #minnowpower

Example of an ad from https://steembottracker.com #minnowpower

In general, the issues of Steem can be attributed to “the lack of unified governance, moderators, and structured accountability creates incidences of significant abuse” Source: Concepts and Criticisms of Steemit.


EOS is in some ways an evolution of Steem, it doesn’t include peer-reviewed content exactly, but it’s worth looking at as a inflation/staking model.

EOS is Dan Larimer’s 3rd large blockchain protocol and the most ambitious one yet. The goal of EOS is to be the blockchain for business… It allows users to build dApps and interact with the state of those dApps using staked EOS tokens to reserve resources, therefore there are no direct transaction fees. The staked tokens can also be used to vote for governance of the chain — Block Producers are elected by staked EOS tokens and, as of a few weeks ago, staked EOS tokens can also vote for Referenda. Instead of transaction fees, the platform is funded by 5% inflation. Let’s take a look at the inflation model.

EOS Inflation Model

The EOS inflation model is fairly simple. Just like all EOS governance, it can be changed by a popular vote through a referendum.

EOS Reward Model

As I have written about previously, the inflation of EOS is divided into 2 pools. One pool goes to pay Block Producers and the other pool goes to an eosio.saving account. As of the writing of this article, the eosio.saving account has accumulated over 26 Million EOS ($72,939,054.30 @ $2.76/EOS). Approximately 6.5 Million EOS has been paid to Block Producers.

Block Producers that have more than .49% of the total voted tokens will receive a minimum of 100 EOS per day. There are approximately 80 elected Block Producers in active and standby positions. The top 21 by vote weight are the ones who actually do the production of the blocks on the EOS chain and they also sign multi-sig transactions to upgrade and change the EOS system contracts which control the chain’s parameters.

There are no payments or rewards for users staking built into the EOS reward model.

Criticisms of EOS

Governance — Issues involve vote buying, whale domination, concerns about centralization and cabals in the top 21. Also, there is a controversial body of governance called ECAF (EOS Core Arbitration Forum) which was designed to resolve disputes between contracts, but got sucked into a lot of hacking / key recovery issues and has been completely dragged through the mud due to issues with the way that the organization has been conducting itself.

Gambling — Most of the on-chain activity involves gambling. There are very few games on EOS that have any traction compared to the many, many gambling apps. The other dApps and users have to compete with the gambling dApps for resources and therefore resources have been very expensive.

Inflation — There are many people who believe that eosio.saving is growing too large, too fast. There are calls by token holders and Block Producers to burn the saving account however it has not gone to referendum vote yet. Currently there is a proposal up for vote to take 1 million EOS tokens and use it to pay for “emergency support” that Block.One is not adequately providing (B1 is not expected nor did B1 promise to provide such support). This proposal is highly contentious and being debated actively on many EOS channels.

Everipedia (IQ)

Everipedia is a blockchain-based wiki site. I know the team personally. They are a group of smart people and they have some really great developers. The project is best described by it’s own whitepaper abstract:

Screen Shot 2019-02-22 at 11.46.45 am.png

The IQ tokens were given out freely to all EOS accounts based on the genesis snapshot of EOS ERC20 tokens. There is a total supply of 10 billion. The IQ tokens are used for voting in governance and for voting on article content. The token itself can be modified by governance in the system, but the balances and databases containing articles cannot be modified by governance.

Everipedia (IQ) Inflation Model

This is a pretty confusing way of saying that the rate of minting new tokens can be set every 30 minute period based on whether or not there were any edits or votes on articles. As of Aug 2018, TPeriod was set to 10 IQ every 30 minutes that the platform is used within that 30 minutes. However, according to their whitepaper excerpt (above) if inflation is set to 5% then the TPeriod = 28538.813 every 30 minutes. The daily minting rate is between 480 and 1369863.024 (or between $1.41408 and ~$4000 per day @ today’s price of $0.002946/IQ). There is also a 0.1% transaction fee, which is deflationary and can be changed to control the token velocity.

Everipedia (IQ) Reward Model

Everipedia gives rewards for curating (voting) and editing articles. The weight of a user’s vote in the system is determined by “Brain Power.” Staking IQ locks up the IQ token for 21 days (non-transferrable) and it gives the user a 1:1 ratio of spendable “Brain Power”(non-tradable). After 21 days, the tokens can be staked again to accrue new “Brain Power.”

Curation and edit rewards are paid to users who do the editing and those who vote for or against the validity of the user’s activity (curating). The “Brain Power” in the equation is based on the amount of staked IQ. If a user wins, they will be rewarded IQ, if the user loses, then their staking period will be increased (this is called a slash).

Criticism of Everipedia (IQ)

There is actually a page for this on the Everipedia site which can be edited by any Everipedia user.

Everipedia was one of the only EOS projects to receive funding from EOS VC before the launch of the EOS Mainnet. The team raised $30 million dollarsand gave out the rest of their tokens for free via an airdrop on EOS. One critique I have heard is that it is unclear how the Everipedia team will make money and/or pay back their investors since Wikipedia does not make any profits and relies on donations.


Karma is a relatively new dApp built on EOS which just recently got approved in the Apple App Store. (Congrats Dallas Rushing!)

The Karma mobile app is an Instagram-like feed of good deeds from around the world. Instead of posting vacation shots to cause FOMO with your friends and followers, Karma incentivizes and rewards people to post videos and photos of people helping people. I have deleted Instagram from my phone, but I do love browsing the content on Karma to see what sort of things people are posting from around the world. The token in Karma must be Powered Up (staked) to vote on content (Up/Down).

Let’s take a high level look at their tokenomics.

Karma Inflation Model

In the Karma inflation model, the rate of annual inflation is calculated daily based on the amount of posts being made. When there are more posts each day, the rate of inflation is higher. At the current price of Karma ($0.000483 / 1 KARMA), this is between $125 and $625 newly minted and distributed tokens per day.

Karma Reward Model

The inflation of Karma is distributed to it’s users divided by 4 pools:

Content Creators — 70%

  • 95% of the 70% goes to video posts.

  • 5% of 70% goes to photo posts.

Staking Rewards — 15%

  • Reward must be “claimed”

  • Staked (Powered Up) tokens are locked for 3 days after unstaking

Engagement — 10%

  • Users who post, comment, or upvote/downvote.

  • (New users can acquire Karma this way by commenting and voting.)

Administration — 5%

  • The Administrator of Karma (Is this Dallas?).

The Karma model for rewards is interesting because it allows users with no Karma the chance to earn Karma just be engaging in the system. AFAIK, users of Karma must have an EOS account which costs ~$0.50 at current EOS prices.

Criticism of Karma

Karma is very new and I’m a big fan of Dallas as well as his project. I don’t have much criticism of Karma, but I do have questions.

I love how scrappy the app is — UI could be better, but I don’t think it’s a total blocker to users. The claim feature is a bit hard to find in the app, is this intentional?

How decentralized is Karma? I’m not sure how the token is used for governance of the Karma platform. Currently, it seems like the “Administrator” of Karma has control to change the inflation, control content, and also change the rewards. I believe I read somewhere that Karma will be launching a DAC, in which case I imagine the Karma token will be used for electing and voting on governance items.

Where is Karma storing the photos and videos? IPFS? AWS? Can a user export their data? I have not dug much into Karma to see what is happening behind the scenes and the app is not open source. I may do a network packet capture later this week to see if I can sniff it out.

Who pays for the EOS transactional resources for users? The Karma app has comments and votes stored on-chain in EOS. Currently, the Karma dApp contract only has 455 EOS staked to it for resource utilization on EOS. I’m not sure how new users without EOS staked in their accounts use the platform, but that might be a bottleneck for usage if users are required to have their own EOS staked to CPU/Net to be able to use the app.


All of the projects outlined here use staking methods to decide who gets the newly minted tokens from inflation. Inflation rates vary greatly from project to project. One aspect of the Everipedia token that I think is smart is that the inflation is tied to the usage of the platform. In Steem, there are Steem Dollars, Steem, and Steem Power — which makes up a complicated system, but having a “stable coin” that is used within the network to reward one-another is very cool.

I believe there are some good non-inflationary reward models that have not been explored yet. This is something I have actively been thinking about… The goal of token system in a social platform is to provide a means for governance, reward, and security against spam, trolling, and sybil attacks.

A properly designed finite token supply system that allows transactional rewards might actually be less prone to whale monopolies. In DPoS systems, it seems quite common that those with the most tokens have the most power and they use that power to get even more tokens.

There are some very exciting experimental economies developing in blockchain that are geared towards rewarding the people who contribute, shape, and participate in new platforms. The internet is a truly amazing conduit for free flowing information, let’s support projects that intend to keep it that way.

BP1 Proxy Show

BP1 Proxy Show

Why Build on EOSIO?

Why Build on EOSIO?