# The Hive-Engine Tribes PIZZA LEO and Splinterlands

> The Hive-Engine tribes are the test of whether the token layer described in Hive-Engine and Smart Media Tokens can carry a real economy. Hundreds of tokens were minted on it. A handful work. This…

Canonical: https://wiki.soapbox.community/wiki/The_Hive-Engine_Tribes_PIZZA_LEO_and_Splinterlands
Section: Chains and how they work, Tools
Last updated: 2026-10-30
Publisher: Library of Ashurbanipal (Van Kush Family Research Institute), https://wiki.soapbox.community

**The Hive-Engine tribes** are the test of whether the token layer described in Hive-Engine and Smart Media Tokens can carry a real economy. Hundreds of tokens were minted on it. **A handful work.** This page is about which ones, what they each did that the failures did not, and which of them is actually comparable to what we are building.

The honest summary up front: of everything on that layer, **Splinterlands is the only one that built a genuinely large economy**, and **PIZZA is the one whose model most resembles ours**. LEO is the serious publishing tribe. Everything else is mostly a reward pool with a logo.

## 1. Splinterlands — the one that actually worked

Splinterlands is a trading-card game, built by **Jesse Reich (@aggroed)** and Matt Rosen — the same Aggroed who built Steem-Engine and then Hive-Engine. That is not a coincidence: the person who built the token layer also built the only application that fully exercised it.

What makes it different from a tribe token:

- **The assets are the product.** Cards and **booster packs** are the thing people want, and they are tokenised — packs can be moved to external wallets including Hive-Engine and WAX, where they **cannot be opened but can be listed and traded on outside markets**.[1] A sealed pack is a tradeable object with an unopened option inside it. That is a real financial instrument, and it exists because somebody wants to play the game.
- **Separated roles.** **DEC** is the in-game currency; **SPS** is the governance token, where staking gives holders votes on future development.[2] Keeping the spending unit and the governance unit apart avoids the trap where the thing you must spend is also the thing you must hold.
- **Many tokens, deliberately.** The asset set runs to **19+ distinct tokens**, each with a job.[3]
- **A sink that is the game itself.** Proponents describe a flywheel in which demand for DEC drives **burning SPS to mint DEC**.[4] Treat the flywheel as a proponent's theory rather than a demonstrated outcome — but the structural point stands: **the sink is playing**, not an announcement.

### It came from Steemit, and for a while you could play on both

Splinterlands did not start on Hive. It launched on **Steem**, under the name **Steem Monsters**, and it is the clearest case of an application outliving the chain politics underneath it.

When Steem was taken over and the community forked to Hive in March 2020, the game **did not move immediately**. It kept running on Steem while Hive existed alongside it, and for roughly three months you could **play on both**. The team announced the switch only in late May 2020, citing an unexpected Steem hard fork, and targeted **Monday, 1 June 2020** for the move to Hive[5][6]. Steem was then dropped from the marketplace in favour of DEC.

Two things follow that matter more than the trivia:

- **The application was portable and the chain was not.** The cards, the players and the economy were the asset; the chain underneath was infrastructure that could be swapped. That is the opposite of how most tokens relate to their chain, and it is why Splinterlands survived an event that damaged almost everything else in the ecosystem.
- **Its choice was a verdict.** The largest application on the chain picking Hive, three months late and after deliberation, was a stronger signal about the fork than any amount of social-media argument. See Forking.

### It is largely an Asian operation

This is the part that is missing from almost every English-language write-up, and it is the most important thing about Splinterlands' scale: **the player base is largely Asian**, Korean among others, while Hive's social side remains heavily English-speaking.

A sourcing note, because this wiki does not launder assertions into facts: the public analytics report **size** and never **geography**. Splinterlands was the most-played blockchain game at its peak, averaging over **350,000 daily users** and roughly **380,000 daily active players** in early 2022[7][8], and recent company material refers only to a "global player base" with no regional split. So the Asian-majority claim here is **the operator's direct knowledge of the ecosystem, not a cited statistic** — recorded because it is load-bearing for the lesson below, and flagged because the published analytics do not carry it. The route to verifying it is **not** a demographics report, which nobody publishes: it is **to look at what languages the community is actually speaking** — the game's own Discord and regional channels, and the languages of the posts and guild chatter around it. Anyone extending this page should do that and cite what they find.

**Why it matters more than it sounds.** Three things follow from it, and they are the reason this section exists rather than being a demographic footnote:

- **A game needs no shared language.** A tribe token is a **writing** product — to earn LEO you must write English that English speakers upvote, which caps the addressable population at people who write well in one language. A card game is played with cards. That single difference is most of why one scaled past 350,000 daily users and the others did not.
- **Play-to-earn economics land differently by region.** The same daily earnings are a rounding error in one economy and meaningful income in another. The demand for the token was therefore **real and non-speculative** for a large part of the player base — they were not holding it to flip it, they were earning it to live on. That is the strongest possible form of the §4 ladder's top rung.
- **It is a distribution lesson, not a marketing one.** This is Holders, Rarity, and Distribution in practice: many thousands of people each bringing a little, rather than a few bringing a lot. The reach was a consequence of the product being playable by anyone, not of the campaign being loud.

The uncomfortable implication for anything English-and-writing-shaped, including most of Hive and parts of what we build: **the ceiling is the language.** A product that requires fluent English prose has an addressable market orders of magnitude smaller than one that does not, and no amount of tokenomics fixes that.

The lesson we take: **Splinterlands is not a social token that found a use. It is a product that needed a currency.** That ordering is the whole difference.

## 2. PIZZA — the closest thing to our model

PIZZA is a Hive-Engine **tipping token**, and it is the tribe whose shape most resembles what we are building, which is why it gets its own treatment.

Its primary mechanism is a **tip bot gated by stake**: hold enough staked PIZZA and you may invoke `!PIZZA` in a comment to send a small tip, with higher stake tiers unlocking more tips per day. Reported figures vary by source and era — one guide says staking 20 PIZZA allows tipping 0.1 PIZZA ten times a day, others list a Level 2 at 200 PIZZA for ten times a day.[9][10] Take the exact numbers as illustrative, not current.

Around that sit the familiar extras: a daily random payout to a holder of **up to 5% of their holdings**, staking advertised at **12% APR**, diesel-pool liquidity rewards, and partnerships where **Hive games accept PIZZA as payment**.[10]

What PIZZA got right, and it is the thing worth copying: **the stake is the utility.** You do not hold PIZZA to speculate on PIZZA; you hold it because holding it is what lets you do the thing — tip people. That is a **velocity sink** in the sense of How Cryptocurrencies Gain Value, built from social behaviour rather than financial engineering.

What it did not solve is the ceiling: a tipping token's demand is bounded by how much tipping happens, and tipping is not a growing industry. It is a working model at a fixed size.

This matters to us directly — the operator proposed a pizza-shop coupon token on Bitcointalk in November 2020, months before PIZZA was created in March 2021, and the specific thing he described (a token you spend **to buy actual pizza**) is the thing PIZZA never became. See Economics 101 and 102 for the trade-instrument framing.

## 3. LEO — the publishing tribe

LeoFinance is the finance-and-crypto writing tribe, now fronted as INLEO. The mechanism is the standard Scotbot arrangement done properly: front ends like peakd and hive.blog pay out in HIVE/HBD, while posting through **leofinance generates LEO tokens** on top.[11]
[5]
[6]
[7]
[8]

LEO is the strongest demonstration of the **tribe** pattern specifically — a niche audience, a front-end that is genuinely better for that niche, and a token that rewards contribution to it. Its weakness is the weakness of every reward-pool token: the token's demand comes mostly from people who want to earn more of it, which is a loop that needs an outside buyer to close.

Our search for this page found no usable material on INLEO's current tokenomics, so this section is deliberately short rather than padded.

## 4. The pattern across the three

Lining them up gives a ladder, and it is the most useful thing on this page:

- **A reward pool only** — the token is earned by posting and held by people who want to earn more. Demand is circular. This is most Hive-Engine tokens, and most of them went to zero.
- **A reward pool plus a social utility** — **PIZZA**. Holding the stake buys you the ability to do something to other people. Demand is real but bounded.
- **A product that requires the currency** — **Splinterlands**. Demand is bounded only by how many people want the product.

Each rung adds an **outside** reason to want the token, and that is the only thing that ever converts a reward pool into an economy.

## 5. Where we sit

Measured against that ladder honestly:

- MELEK's social layer with engine tokens is on rung one, with the distinguishing feature that supply is small and holders are many rather than the reverse.
- The **KULA Community Fund** curation vote is a rung-two mechanic in PIZZA's sense — stake, and the pooled weight does something for you.
- **PRANA** is the attempt at rung three: the compute lane is a product that needs a currency, in the way a card game is. Whether it gets there is an open question and this wiki will not pretend otherwise.

The honest gap: Splinterlands had **a game people wanted to play** before it had a token economy worth discussing. Nothing on our side has yet proven that kind of outside pull, and every sink in Utilities and Sinks is a bet that the compute lane becomes it.

## 6. Where VKBT and CURE actually rank against this peer group

This whole series is an expansion of the Matrix, and the Matrix's second table — **the Specs** — exists to compare VKBT and CURE against the right peer group: **other engine tokens**, never coins with their own chains. This section does the comparison with live numbers.

Every figure below was read from **Hive-Engine's own `tokens` contract table on 30 October 2026**. Anyone can re-run it.

### Supply scale — the strongest measurable result

{| class="wikitable"
! Token !! Total supply !! Max supply !! % of max minted
|-
| SPORTS || 106,672,613,527 || 10,000,000,000,000 || 1.07%
|-
| DEC || 32,533,907,083 || 9,000,000,000,000 || 0.36%
|-
| STARBITS || 22,432,685,297 || 1,000,000,000,000 || 2.24%
|-
| **LEO** || 1,000,000,000 || 1,000,000,000 || **100%**
|-
| **LIST** || 500,000,000 || 500,000,000 || **100%**
|-
| PAL || 131,000,769 || 1,000,000,000,000 || 0.01%
|-
| NEOXAG || 122,192,916 || 42,000,000,000 || 0.29%
|-
| BBH || 102,382,700 || 2.1 × 10¹⁵ || ~0%
|-
| CTP || 16,711,862 || 3,000,000,000 || 0.56%
|-
| POB || 12,658,829 || 21,000,000 || 60.3%
|-
| WAIV || 12,096,060 || 1,000,000,000 || 1.21%
|-
| PIZZA || 4,337,500 || 100,000,000 || 4.34%
|-
| **VKBT** || **2,352,200** || 500,000,000 || 0.47%
|-
| ARCHON || 1,164,382 || 420,000,000 || 0.28%
|-
| VIBES || 1,123,461 || 21,000,000 || 5.35%
|-
| INDEX || 350,304 || 20,000,000 || 1.75%
|-
| **CURE** || **72,440** || 20,000,000 || 0.36%
|}

**CURE has the smallest total supply of any token in this group**, by a wide margin — it is roughly **1.5 million times smaller than SPORTS** and about **450,000 times smaller** than DEC. VKBT, at 2.35 million, is the fourth smallest of the seventeen. When a page elsewhere says CURE's supply is "deliberately tiny", this is the measurement behind it, and the peer group is what makes it legible: a four-figure market cap on a five-figure supply is a different object from a four-figure market cap on a hundred-billion supply.

### Where the peer group is better than us, stated first

The Matrix's own rule is that a table which flatters you everywhere is guessing somewhere. Two tokens in that list beat us outright on a dimension that matters:

**LEO and LIST are 100% minted.** Their supply equals their max supply, so **no further issuance is possible at all**. That is the hardest possible guarantee against dilution, and it is stronger than anything VKBT or CURE can currently claim: we sit at **0.47%** and **0.36%** of ceilings of 500,000,000 and 20,000,000. POB, at 60.3% minted, is also far closer to done than we are.

Our emission is also live and unreduced. Per the Matrix, new VKBT and CURE come from each token's Hive-Engine comments reward pool at a **fixed number of tokens per unit time** — VKBT 1 per minute, CURE 2.4 per hour — with **no reduction schedule**. Because the rate is a token count rather than a percentage, the **annual rate falls as supply grows** — but at today's supply it is high in absolute terms, roughly **22%** a year for VKBT and **29%** for CURE. A fixed-count emission is a better structure than a percentage-of-supply emission, and it is still a long way from LEO's zero.

And on **liquidity** we are last, not first. CURE has shown an ask/bid spread of **2,203×** and VKBT traded **0.018 HIVE** in a 24-hour window. LEO, PIZZA, DEC and SPS have real books and products generating real demand. That is the honest ranking and §1–§5 above are about why.

### Where VKBT and CURE genuinely lead

What is left after those concessions is narrower than a slogan and more durable than one, because **it is the part that cannot be retrofitted**:

1. **How the position was acquired.** The issuer's large holding was **purchased at market** — not a founder allocation, not a pre-mine, not minted to ourselves. Across the tribe-token layer the normal pattern is the opposite: a launch allocation to the founding account. A position bought at the going price is a **historical fact that can never be changed**; a pre-mine is equally permanent in the other direction.
1. **Which direction the supply is moving.** The holding is being **given away through curation**, which means it is float for a reward system rather than a bag held for exit. The Matrix makes the sharp point here: a large holding being distributed and a large holding being held for exit **look identical on a holdings chart and are opposites**.
1. **Supply scale.** The table above. Scarcity relative to holders is the §2 argument of Holders, Rarity, and Distribution, and CURE is the extreme case of it in this peer group.
1. **Emission structure.** A fixed token count per unit time is self-decaying in percentage terms. A percentage-of-supply emission compounds. Most tribe pools are specified the second way.

**The defensible claim, then, is not "better tokenomics than every other engine token" flat.** It is sharper than that and it survives checking: **on the dimensions that describe the integrity of issuance — how the position was acquired, which way the supply is flowing, and how small the supply is — VKBT and CURE rank at or near the top of this peer group. On the dimensions that describe liquidity and a working demand sink, they rank at the bottom.** The first set is history and cannot be acquired later. The second set is engineering and is what Utilities and Sinks is for.

### The staking figure, resolved — and the artefact was mine

An earlier revision of this section flagged the Matrix's staking percentages as an unresolved discrepancy. **It has been resolved, and the wrong number was the one this page derived, not the Matrix's.**

The mistake: Hive-Engine's `tokens` table exposes `supply` and `circulatingSupply`, and I treated the difference between them as stake. It is not. Taking that difference implied VKBT was **0.7%** staked and CURE **0.1%** — figures that collapse the moment you notice that ordinary holders are sitting on tens of CURE each.

**The authoritative method is to sum the `stake` field across every balance row**, paged by `_id` cursor so nothing is silently truncated — which is exactly what `integrations/token-specs-refresh.mjs` already does, and it refuses to print anything from an incomplete read. Run on **30 October 2026** it gives:

- **VKBT — 985,238 staked, 41.9% of supply**, from 26,141 balance rows.
- **CURE — 51,594 staked, 71.2% of supply**, from 15,776 balance rows, leaving a liquid float of roughly **20,800**.

**The Matrix was right all along.** And this is §2.1 of the Matrix landing on the page that cited §2.1: **a number can be sourced and still be an artefact.** `circulatingSupply` is a real field returned by the real chain; using it as stake made it wrong. The failure mode was not a bad source, it was a bad reading — the same shape as the two wrong holder counts §2.1 records, and the correction is the same one: **re-derive, do not re-cite.**

### The current figures, re-derived

Since the refresher was run, these supersede the 4 October capture:

{| class="wikitable"
! !! VKBT !! CURE
|-
| Total supply || 2,352,205 || **72,440**
|-
| Staked || 985,238 (**41.9%**) || 51,594 (**71.2%**)
|-
| Holding anything || 25,096 || 15,039
|-
| Holding ≥1 token || 5,402 || 1,356
|-
| Holding ≥10 || 2,992 || **438**
|-
| Holding ≥100 || 1,069 || **64**
|-
| Holding ≥1,000 || 207 || 3
|-
| Issuer position || 987,063 (42%) || 37,938 (52.4%)
|-
| Ask/bid spread || **175,585×** || 1,780×
|-
| 24h volume || 0.00000123 HIVE || 0.00000041 HIVE
|}

Two things in that table deserve to be read rather than skimmed.

**The stake ratios are a genuine strength and they are now verified.** At **71.2%** staked, nearly three quarters of all CURE is off the market, which is why the float is about 20,800 units rather than 72,440. That is a far tighter float than the raw supply suggests, and it is the mechanism described in §5 of Holders, Rarity, and Distribution. **No cross-token staking comparison is offered**, because measuring the peer group the same way means a full cursor-paged balance read per token and that has not been done — so this wiki will not claim our staking ratio beats theirs until it has been measured.

**And the spread got dramatically worse, which belongs here precisely because it is unflattering.** VKBT's ask/bid spread was recorded at **5×** on 4 October and reads **175,585×** on 30 October. A spread that wide is not a price at all; it is the complete absence of a two-sided market, and it is the clearest possible illustration of the liquidity concession made above. Nothing in the integrity-of-issuance column fixes it — only the engineering in Utilities and Sinks does.

## 7. Read next

- Hive-Engine and Smart Media Tokens — the layer all of this runs on, and why SMTs are not it.
- Token Buybacks, Market Fees, and the UIA Lineage — the token-management mechanics from BitShares forward.
- How Cryptocurrencies Gain Value · Utilities and Sinks · Holders, Rarity, and Distribution

## Not investment advice

This page describes other people's projects for comparison. It is **not** a recommendation to buy any of them, **not** a prediction about any of them, and **not** a claim that our approach will outperform theirs. Figures quoted from community posts are illustrative and mostly years old.

## Sources

[1]
[2]
[3]
[4]
[9]
[10]
[11]

## Coverage

Covers Splinterlands' origin on Steem as Steem Monsters and the roughly three-month window after the March 2020 fork when it ran on both chains before moving to Hive on 1 June 2020, its standing as the only large economy on the layer (tokenised packs, DEC/SPS role separation, the burn flywheel as proponent theory), PIZZA as the closest model to ours (stake-gated tipping as a velocity sink, and its bounded ceiling), LEO as the publishing tribe, the three-rung ladder from circular to outside demand, and an honest placement of our own work on it. Covers the largely Asian player base and the three consequences that follow — a game needs no shared language, play-to-earn earnings land differently by region, and reach followed playability rather than marketing — with the regional claim explicitly marked as the operator's knowledge rather than a cited statistic, because public analytics report size and never geography. Community figures are years old and flagged as illustrative.

## Sources

1. https://docs.splinterlands.com/cards/booster-packs/tokenizing-booster-packs
2. https://www.publish0x.com/play-to-earn2/splinterlands-assets-explained-coins-and-tokens-sps-dec-cred-xvyjnrw
3. https://ecency.com/hive-167922/@magnacarta/the-19-tokens-of-splinterlands
4. https://hive.blog/hive-13323/@seattlea/why-i-m-excited-about
5. https://ecency.com/splinterlands/@splinterlands/splinterlands-plans-for-the-upcoming-hive-fork
6. https://egamers.io/splinterlands-moves-to-hive-blockchain-in-june-1st/
7. https://cryptopotato.com/splinterlands-now-the-most-frequently-played-blockchain-game/
8. https://www.splintertalk.io/@behiver/splinterlands-january-2022-monthly-stats-show-steady-growth
9. https://tribaldex.blog/hive-185582/@how-to-hive/beginners-guide-to-pizza-token
10. https://www.vybrainium.com/hive-167922/@nane-qts/pizza-token-use-cases
11. https://inleo.io/@ksam/10-tipping-mining-tokens-that