"I'd love to come, but I just can't afford it right now."
We've heard this many times. Conference tickets. Event accommodations. NFT mints. Exclusive access. The costs add up, and the people with the most to contribute often have the least to spend.
This is the distribution problem at the heart of web3. How do you allocate scarce resources fairly? How do you give opportunities to people who deserve them, not just people who can afford them or got lucky?
One of the default answers has been raffles. Random selection. Everyone gets a chance. But after years of watching web3 raffles play out, we've learned something uncomfortable: randomness isn't the same as fairness.
This post proposes a different primitive: conviction-weighted selection, where communities decide who wins through stake-backed evaluation.
Before we fix raffles, let's understand what they're trying to do.
A raffle is a distribution mechanism for scarce resources. You have something valuable (tickets, access, prizes) and more people want it than can have it. The raffle decides who gets it.
Traditional raffles usually optimize for one thing: equal chance. Buy a ticket, enter the pool, random selection picks winners. The appeal is simplicity and perceived fairness. Everyone has the same odds per entry.
But equal chance isn't the same as fair distribution. Consider what a random raffle actually selects for:
What's Measured | What's NOT Measure |
|---|---|
Willingness to enter | Merit or contribution |
Ability to afford entries | Need or deservingness |
Luck | Community value |
Bot sophistication | Authentic interest |
Random selection treats all entries as equivalent. A bot farm with 1,000 wallets has 1,000x the chance of a genuine community member with one wallet. A whale who can afford unlimited entries dominates someone who scraped together funds for one shot.
The mechanism is simple, but the outcomes often aren't fair at all.
The common thread across existing mechanisms: they don't measure what matters.
If you're giving away conference tickets, you probably want them to go to people who will engage meaningfully, contribute to conversations, and become lasting community members. None of the mechanisms above select for that.
If you're distributing access to an exclusive community, you probably want members who will participate actively and add value. Random selection doesn't care about that.
If you're allocating grants or scholarships, you want recipients who will use the resources well. Lottery tickets don't measure capability.
The mechanisms optimize for what's easy to measure (wallet count, token balance, task completion, speed) rather than what actually matters (merit, need, contribution potential, community fit).
And beyond measurement, there's a deeper problem: no consequence for bad selection.
In most raffle mechanisms, voters or selectors have nothing at stake. A committee member who picks their friend faces no penalty. A social task that gets gamed costs the gamer nothing. There's no feedback loop that improves selection quality over time.
Without consequence, there's no signal. Just noise dressed up as fairness.
What if we designed a raffle mechanism from first principles?
Requirements:
Open participation (anyone can enter)
Community decides winners (not random, not committee)
Merit can surface (the "right" person can win)
Evaluators have skin in the game (consequence for bad judgment)
Curation work is compensated (evaluation isn't free labor)
Sybil-resistant (gaming is expensive)
This is what we built at Incented. The core mechanism is conviction voting: stake-weighted community evaluation where correct judgment earns rewards and incorrect judgment has a cost.
Here's how it maps to raffle primitives:
Raffle Element | Incented Equivalent |
|---|---|
Entry | Application/submission |
Ticket price | Free to apply (barrier is effort, not money) |
Random draw | Community vote (stake-weighted) |
Winner selection | Highest net votes / quorum decision |
Prize | Award pool distributed to winners |
Reward Curation | Voting pool rewards evaluators |
Entries are evaluated, not drawn. Instead of random selection, the community reviews applications and stakes tokens on who should win. This creates actual signal about merit.
Voting has consequence. Stake tokens FOR a submission you believe should win. If it wins, earn rewards. If it loses, lose a percentage of your stake (the "slash"). This filters out casual voting and creates genuine evaluation.
AGAINST votes matter. You can also stake AGAINST submissions you believe shouldn't win. If they lose, you earn. If they win, you lose stake. This creates a quality floor, not just a ceiling. Bad applications get filtered by community signal.
Evaluators get paid. The voting pool compensates people who do the work of assessment. Curation is labor. The mechanism recognizes that.
Sybils are expensive. Every fake account needs real tokens to stake. Gaming requires capital at risk. The economics flip: attacking costs money, defending earns money.
Let's walk through the mechanics.
Prize: 1 conference ticket ($500 value)
Voting token: Community token (or SEEDS)
Award pool: The ticket (or equivalent value)
Voting pool: 100 tokens to reward evaluators
Slash rate: 10% (lose 10% of stake if you vote incorrectly)
Anyone can apply. They submit why they deserve the ticket: what they'll contribute, their background, what they want to get from the event. The barrier is effort and authenticity, not money.
Applicants share their submissions. "I applied for the DevConnect ticket, here's my application." Their networks can see who they are and decide whether to support them.
This creates discovery. The community learns about people they didn't know. "Wait, you're building what? I want to meet you there."
Token holders stake on applications:
FOR: "I believe this person should win"
AGAINST: "I believe this person should not win"
Staking isn't free signaling. Your tokens are locked until settlement. If you're wrong, you lose a percentage.
Phase 4: Settlement
Program closes. Net votes calculated. Highest net votes wins.
The winner receives the award. Correct evaluators earn rewards. Incorrect evaluators lose stake.
The Result:
The person who wins isn't random. They're the person the community, with skin in the game, decided was most deserving. Voters who did their homework earned tokens. Voters who staked carelessly lost some.
The mechanism produced signal, not noise.
Incented programs are configurable. Different raffle types need different settings.
Setting | Use Case |
|---|---|
Top 1 | One grand prize (conference ticket, room) |
Top X | Multiple prizes (10 scholarships, 5 access passes) |
Quorum | Everyone above threshold wins (quality bar, not competition) |
Duration | Use Case |
|---|---|
Short (3-5 days) | Quick giveaways, time-sensitive prizes |
Medium (7-14 days) | Standard programs, enough time for evaluation |
Long (30+ days) | Major grants, high-stakes decisions |
Rate | Effect |
|---|---|
Low (5%) | Encourages participation, lower consequence |
Medium (10-15%) | Balanced stakes, meaningful but not punishing |
High (20%+) | High-stakes decisions, strong filter for serious evaluators |
Requirement | Use Case |
|---|---|
Open text | Simple giveaways, low barrier |
Structured form | Grants needing specific info |
Video required | High-touch selection, personality matters |
Portfolio links | Skill-based selection |
We're running this mechanism ourselves. Carrot Castle is a community house during major crypto events. Rooms cost $750-1,150. A real barrier for many.
Community members kept saying: "I'd love to come but can't afford it."
So we created two programs:
Castle Quest Denver (ETH Denver, Feb 15-22)
1 private room ($1,150 value) sponsored by Incented
Community votes using SEEDS tokens
Winner receives CARROT CASTLE COIN redeemable for the room
Castle Quest Boulder (Boulder Week, Feb 12-16)
1 private room ($750 value) sponsored by Artizen
Community votes using $ART tokens
Same redemption model
The mechanism is identical. The difference is which community participates. Denver draws from the Incented community. Boulder draws from Artizen's creative funding community. Both help each other grow.
Why this works for us:
Transparent allocation. Sponsors (Incented, Artizen) can verify the community decided, not a committee.
Community engagement. Applications create content. Voting creates participation. The giveaway becomes an event, not just a transaction.
Right person wins. Not random. Not whoever knows the organizers. The person the community signals should be there.
Voters earn. Community members who evaluate thoughtfully get compensated. Curation labor is recognized.
Discovery. We'll meet applicants we didn't know. The Castle gets more interesting people because the selection surfaced them.
Here's what makes conviction-weighted raffles genuinely different from random giveaways:
The mechanism creates mutual support.
Imagine you're a community member who can afford tokens but not the room. You participate by voting thoughtfully, earn from the voting pool, and help someone who needs it more get their spot.
Imagine you're someone who can't afford either. Apply anyway. If your community believes you deserve it, they rally behind you. Their stakes become your path to the prize.
The people with resources help select. The people who need resources can win.
Both sides are rewarded. Winners get the prize. Voters get compensated for judgment. The community strengthens because people supported each other through mechanism, not charity.
This is what coordination infrastructure should do. Not extract value, but create systems where communities support each other with aligned incentives.
Conviction-weighted selection isn't always the answer.
Use conviction voting when:
The "right" winner matters (not just any winner)
You want community engagement, not just distribution
Merit, need, or contribution should influence selection
You're willing to reward evaluators for their work
You want Sybil resistance through economic cost
Use random when:
Any winner is fine
Speed and simplicity trump precision
Stakes are low enough that gaming doesn't matter
You can't fund a voting pool
The pattern works for any community giveaway:
Element | What You Need |
|---|---|
Prize | Something valuable: tickets, access, merch, services |
Voting Token | Your community's token |
Award Pool | The prize itself or equivalent value |
Voting Pool | Budget to reward evaluators (we suggest 5-10% of prize value) |
Timeline | Application period + voting period (7-14 days typical) |
Setup steps:
Design your program (prize, tokens, duration, requirements)
Create the program on Incented
Announce and open applications
Promote voting to token holders
Settlement distributes prize and voter rewards automatically
What you get:
Transparent, verifiable selection
Community engagement (applications, voting, discussion)
Sybil resistance through staking
Compensated curation labor
Signal about who your community values
The same primitive that powers fair raffles also solves harder coordination problems:
Use Case | How It Works |
|---|---|
Grants programs | Research proposals evaluated by the community |
Bounties | Task completion validated by conviction voting |
Governance | Proposals requiring commitment, not drive-by votes |
Content curation | Quality ranking through stake-backed evaluation |
Scholarships | Recipients chosen by community with skin in the game |
The building blocks are flexible. Top X winners or quorum-based. Fixed prizes or proportional splits. Single rounds or recurring cycles.
The primitive works wherever you need fair distribution with community signal.
We started Incented with a simple thesis: incentive alignment produces better coordination than top-down control.
Raffles are a perfect test case. The traditional approach is either random (no signal) or committee-based (gatekeepers). Neither produces the outcome communities actually want: the right person getting the right opportunity.
Conviction voting offers a third path. Open participation. Community decision. Skin in the game. Consequence for bad judgment. Rewards for good curation.
The stick makes the carrot meaningful. When incorrect votes cost tokens, people vote carefully. When curation work is compensated, people do the work. When merit can surface through community signal, the right person wins.
Castle Quests Boulder & Denver open for applications January 19th. If you can't afford a room but you'd be valuable at the Castle, apply. Let your community decide.
If you have tokens and want to help select who represents your community, vote thoughtfully and earn.
And if you're running a community that wants to give back without gatekeepers, consider what conviction-weighted selection could look like for your next giveaway.
The mechanism exists. The infrastructure is live. The community decides who wins.
Apply for Castle Quest:
Denver: carrotcastle.xyz
Boulder: carrotcastle.xyz
Learn the mechanics:
Run your own conviction raffle:
Get voting tokens:
SEEDS: Juicebox on Base
$ART: Artizen Endowment
The Castle awaits. Let your community decide who joins.
Subscribe to Noises from the Rabbit Hole
<100 subscribers
Noises from the Rabbit HoleBeyond InfoFi
From Attention Farming to Conviction Coordination
Noises from the Rabbit HoleIncented'25 - From Validation to Velocity
A year of building, shipping, and proving that coordination doesn’t have to be complicated.
Noises from the Rabbit HoleThese Carrots Ain't Gonna Eat Themselves
We're Sofa King Tired of All This Innovation
Support Noises from the Rabbit Hole
Noises from the Rabbit HoleRaffles in Web3: Can we do better than Random Chance?
"I'd love to come, but I just can't afford it right now."
We've heard this many times. Conference tickets. Event accommodations. NFT mints. Exclusive access. The costs add up, and the people with the most to contribute often have the least to spend.
This is the distribution problem at the heart of web3. How do you allocate scarce resources fairly? How do you give opportunities to people who deserve them, not just people who can afford them or got lucky?
One of the default answers has been raffles. Random selection. Everyone gets a chance. But after years of watching web3 raffles play out, we've learned something uncomfortable: randomness isn't the same as fairness.
This post proposes a different primitive: conviction-weighted selection, where communities decide who wins through stake-backed evaluation.
Before we fix raffles, let's understand what they're trying to do.
A raffle is a distribution mechanism for scarce resources. You have something valuable (tickets, access, prizes) and more people want it than can have it. The raffle decides who gets it.
Traditional raffles usually optimize for one thing: equal chance. Buy a ticket, enter the pool, random selection picks winners. The appeal is simplicity and perceived fairness. Everyone has the same odds per entry.
But equal chance isn't the same as fair distribution. Consider what a random raffle actually selects for:
What's Measured | What's NOT Measure |
|---|---|
Willingness to enter | Merit or contribution |
Ability to afford entries | Need or deservingness |
Luck | Community value |
Bot sophistication | Authentic interest |
Random selection treats all entries as equivalent. A bot farm with 1,000 wallets has 1,000x the chance of a genuine community member with one wallet. A whale who can afford unlimited entries dominates someone who scraped together funds for one shot.
The mechanism is simple, but the outcomes often aren't fair at all.
The common thread across existing mechanisms: they don't measure what matters.
If you're giving away conference tickets, you probably want them to go to people who will engage meaningfully, contribute to conversations, and become lasting community members. None of the mechanisms above select for that.
If you're distributing access to an exclusive community, you probably want members who will participate actively and add value. Random selection doesn't care about that.
If you're allocating grants or scholarships, you want recipients who will use the resources well. Lottery tickets don't measure capability.
The mechanisms optimize for what's easy to measure (wallet count, token balance, task completion, speed) rather than what actually matters (merit, need, contribution potential, community fit).
And beyond measurement, there's a deeper problem: no consequence for bad selection.
In most raffle mechanisms, voters or selectors have nothing at stake. A committee member who picks their friend faces no penalty. A social task that gets gamed costs the gamer nothing. There's no feedback loop that improves selection quality over time.
Without consequence, there's no signal. Just noise dressed up as fairness.
What if we designed a raffle mechanism from first principles?
Requirements:
Open participation (anyone can enter)
Community decides winners (not random, not committee)
Merit can surface (the "right" person can win)
Evaluators have skin in the game (consequence for bad judgment)
Curation work is compensated (evaluation isn't free labor)
Sybil-resistant (gaming is expensive)
This is what we built at Incented. The core mechanism is conviction voting: stake-weighted community evaluation where correct judgment earns rewards and incorrect judgment has a cost.
Here's how it maps to raffle primitives:
Raffle Element | Incented Equivalent |
|---|---|
Entry | Application/submission |
Ticket price | Free to apply (barrier is effort, not money) |
Random draw | Community vote (stake-weighted) |
Winner selection | Highest net votes / quorum decision |
Prize | Award pool distributed to winners |
Reward Curation | Voting pool rewards evaluators |
Entries are evaluated, not drawn. Instead of random selection, the community reviews applications and stakes tokens on who should win. This creates actual signal about merit.
Voting has consequence. Stake tokens FOR a submission you believe should win. If it wins, earn rewards. If it loses, lose a percentage of your stake (the "slash"). This filters out casual voting and creates genuine evaluation.
AGAINST votes matter. You can also stake AGAINST submissions you believe shouldn't win. If they lose, you earn. If they win, you lose stake. This creates a quality floor, not just a ceiling. Bad applications get filtered by community signal.
Evaluators get paid. The voting pool compensates people who do the work of assessment. Curation is labor. The mechanism recognizes that.
Sybils are expensive. Every fake account needs real tokens to stake. Gaming requires capital at risk. The economics flip: attacking costs money, defending earns money.
Let's walk through the mechanics.
Prize: 1 conference ticket ($500 value)
Voting token: Community token (or SEEDS)
Award pool: The ticket (or equivalent value)
Voting pool: 100 tokens to reward evaluators
Slash rate: 10% (lose 10% of stake if you vote incorrectly)
Anyone can apply. They submit why they deserve the ticket: what they'll contribute, their background, what they want to get from the event. The barrier is effort and authenticity, not money.
Applicants share their submissions. "I applied for the DevConnect ticket, here's my application." Their networks can see who they are and decide whether to support them.
This creates discovery. The community learns about people they didn't know. "Wait, you're building what? I want to meet you there."
Token holders stake on applications:
FOR: "I believe this person should win"
AGAINST: "I believe this person should not win"
Staking isn't free signaling. Your tokens are locked until settlement. If you're wrong, you lose a percentage.
Phase 4: Settlement
Program closes. Net votes calculated. Highest net votes wins.
The winner receives the award. Correct evaluators earn rewards. Incorrect evaluators lose stake.
The Result:
The person who wins isn't random. They're the person the community, with skin in the game, decided was most deserving. Voters who did their homework earned tokens. Voters who staked carelessly lost some.
The mechanism produced signal, not noise.
Incented programs are configurable. Different raffle types need different settings.
Setting | Use Case |
|---|---|
Top 1 | One grand prize (conference ticket, room) |
Top X | Multiple prizes (10 scholarships, 5 access passes) |
Quorum | Everyone above threshold wins (quality bar, not competition) |
Duration | Use Case |
|---|---|
Short (3-5 days) | Quick giveaways, time-sensitive prizes |
Medium (7-14 days) | Standard programs, enough time for evaluation |
Long (30+ days) | Major grants, high-stakes decisions |
Rate | Effect |
|---|---|
Low (5%) | Encourages participation, lower consequence |
Medium (10-15%) | Balanced stakes, meaningful but not punishing |
High (20%+) | High-stakes decisions, strong filter for serious evaluators |
Requirement | Use Case |
|---|---|
Open text | Simple giveaways, low barrier |
Structured form | Grants needing specific info |
Video required | High-touch selection, personality matters |
Portfolio links | Skill-based selection |
We're running this mechanism ourselves. Carrot Castle is a community house during major crypto events. Rooms cost $750-1,150. A real barrier for many.
Community members kept saying: "I'd love to come but can't afford it."
So we created two programs:
Castle Quest Denver (ETH Denver, Feb 15-22)
1 private room ($1,150 value) sponsored by Incented
Community votes using SEEDS tokens
Winner receives CARROT CASTLE COIN redeemable for the room
Castle Quest Boulder (Boulder Week, Feb 12-16)
1 private room ($750 value) sponsored by Artizen
Community votes using $ART tokens
Same redemption model
The mechanism is identical. The difference is which community participates. Denver draws from the Incented community. Boulder draws from Artizen's creative funding community. Both help each other grow.
Why this works for us:
Transparent allocation. Sponsors (Incented, Artizen) can verify the community decided, not a committee.
Community engagement. Applications create content. Voting creates participation. The giveaway becomes an event, not just a transaction.
Right person wins. Not random. Not whoever knows the organizers. The person the community signals should be there.
Voters earn. Community members who evaluate thoughtfully get compensated. Curation labor is recognized.
Discovery. We'll meet applicants we didn't know. The Castle gets more interesting people because the selection surfaced them.
Here's what makes conviction-weighted raffles genuinely different from random giveaways:
The mechanism creates mutual support.
Imagine you're a community member who can afford tokens but not the room. You participate by voting thoughtfully, earn from the voting pool, and help someone who needs it more get their spot.
Imagine you're someone who can't afford either. Apply anyway. If your community believes you deserve it, they rally behind you. Their stakes become your path to the prize.
The people with resources help select. The people who need resources can win.
Both sides are rewarded. Winners get the prize. Voters get compensated for judgment. The community strengthens because people supported each other through mechanism, not charity.
This is what coordination infrastructure should do. Not extract value, but create systems where communities support each other with aligned incentives.
Conviction-weighted selection isn't always the answer.
Use conviction voting when:
The "right" winner matters (not just any winner)
You want community engagement, not just distribution
Merit, need, or contribution should influence selection
You're willing to reward evaluators for their work
You want Sybil resistance through economic cost
Use random when:
Any winner is fine
Speed and simplicity trump precision
Stakes are low enough that gaming doesn't matter
You can't fund a voting pool
The pattern works for any community giveaway:
Element | What You Need |
|---|---|
Prize | Something valuable: tickets, access, merch, services |
Voting Token | Your community's token |
Award Pool | The prize itself or equivalent value |
Voting Pool | Budget to reward evaluators (we suggest 5-10% of prize value) |
Timeline | Application period + voting period (7-14 days typical) |
Setup steps:
Design your program (prize, tokens, duration, requirements)
Create the program on Incented
Announce and open applications
Promote voting to token holders
Settlement distributes prize and voter rewards automatically
What you get:
Transparent, verifiable selection
Community engagement (applications, voting, discussion)
Sybil resistance through staking
Compensated curation labor
Signal about who your community values
The same primitive that powers fair raffles also solves harder coordination problems:
Use Case | How It Works |
|---|---|
Grants programs | Research proposals evaluated by the community |
Bounties | Task completion validated by conviction voting |
Governance | Proposals requiring commitment, not drive-by votes |
Content curation | Quality ranking through stake-backed evaluation |
Scholarships | Recipients chosen by community with skin in the game |
The building blocks are flexible. Top X winners or quorum-based. Fixed prizes or proportional splits. Single rounds or recurring cycles.
The primitive works wherever you need fair distribution with community signal.
We started Incented with a simple thesis: incentive alignment produces better coordination than top-down control.
Raffles are a perfect test case. The traditional approach is either random (no signal) or committee-based (gatekeepers). Neither produces the outcome communities actually want: the right person getting the right opportunity.
Conviction voting offers a third path. Open participation. Community decision. Skin in the game. Consequence for bad judgment. Rewards for good curation.
The stick makes the carrot meaningful. When incorrect votes cost tokens, people vote carefully. When curation work is compensated, people do the work. When merit can surface through community signal, the right person wins.
Castle Quests Boulder & Denver open for applications January 19th. If you can't afford a room but you'd be valuable at the Castle, apply. Let your community decide.
If you have tokens and want to help select who represents your community, vote thoughtfully and earn.
And if you're running a community that wants to give back without gatekeepers, consider what conviction-weighted selection could look like for your next giveaway.
The mechanism exists. The infrastructure is live. The community decides who wins.
Apply for Castle Quest:
Denver: carrotcastle.xyz
Boulder: carrotcastle.xyz
Learn the mechanics:
Run your own conviction raffle:
Get voting tokens:
SEEDS: Juicebox on Base
$ART: Artizen Endowment
The Castle awaits. Let your community decide who joins.
Subscribe to Noises from the Rabbit Hole
<100 subscribers
Support Noises from the Rabbit Hole
Noises from the Rabbit HoleRaffles in Web3: Can we do better than Random Chance?
Noises from the Rabbit HoleBeyond InfoFi
From Attention Farming to Conviction Coordination
Noises from the Rabbit HoleIncented'25 - From Validation to Velocity
A year of building, shipping, and proving that coordination doesn’t have to be complicated.
Noises from the Rabbit HoleThese Carrots Ain't Gonna Eat Themselves
We're Sofa King Tired of All This Innovation
Share Dialog
Share Dialog
IamSvenH
IamSvenH
2 comments
Such a cool use case and a gift to the community!
Thank you. :)