# Skepsis vs binary markets
Source: https://docs.skepsis.live/comparisons/vs-binary-markets
How Skepsis differs from yes/no prediction markets like Polymarket.
Binary markets price whether an event happens. Skepsis prices how much, on a continuous range. Different question, different payoff.
| | Binary market | Skepsis |
| --------------- | ---------------------------------------- | --------------------------------------------- |
| Question | Will X happen (yes/no) | Where does the number land (range) |
| Payoff | Fixed, on direction | Scales with precision: N / k at uniform prior |
| A near miss | Wins the same as a blowout | Pays only for the band that hits |
| Range exposure | Stack many binaries; liquidity fragments | One curve, shared liquidity |
| One-of-N choice | N separate yes/no books | One market, N options, shared depth |
| Hedging fit | Basis risk vs real exposure | Hedge the band the exposure lives in |
| Resolution | Varies, some discretionary | Chainlink numerical feed, time-gated |
| Listing | Often permissionless | Curated |
## When binary is the right tool
If your view really is binary, will this happen or not, a binary market fits. Skepsis is for outcomes that are numbers, where direction is not the whole story and magnitude is the point.
# FAQ
Source: https://docs.skepsis.live/faq
Common questions about trading, payouts, liquidity, and access.
A continuous-outcome prediction market. You price the range a number lands in, not a yes or no. A tighter range pays more, because it is harder to hit.
Arbitrum Sepolia. Trading uses test USDC. No real funds. Mainnet contracts exist; they are not deployed.
With email or Google. A wallet is created for you, or connect your own. See the [quick start](/overview/quick-start).
Payoff scales with precision. At a uniform prior, payoff = N / k for k of N buckets, so a 1-bucket pick on a 10-bucket market pays 10x. The live number is shown before you confirm.
Yes. Close a position at the current price any time before resolution.
From a named Chainlink feed at a set time. Numerical and mechanical. See [trust and resolution](/how-it-works/trust-and-resolution).
One ERC-4626 vault seeds every market and recycles capital on resolution. See [the vault](/for-lps/vault-overview).
Alpha decay thins the book as a market nears resolution, defending liquidity providers against late-information traders. See [alpha decay](/how-it-works/alpha-decay).
Private beta, access in waves. Enter your code in the app, or request access at [skepsis.live](https://skepsis.live).
Admins never DM first and never ask for your seed phrase or private keys. This is testnet; do not use real funds.
# Risks and returns
Source: https://docs.skepsis.live/for-lps/risks-and-returns
What liquidity providers earn, and the bounded risk they take.
## What you earn
The vault earns the spread the LMSR captures on trading, plus fees, across every market it backs. One deposit, exposure across all markets, capital recycled on resolution. See [the vault](/for-lps/vault-overview).
## The risk
The vault is the counterparty to traders. On resolution it pays winning shares from the over-collateralized pool. If informed trading moves a market against the book, the vault can take a loss on that market. This is the risk you are paid for.
## What bounds it
* Solvency: LMSR keeps the pool at or above the maximum payout at all times.
* Bounded loss: the alpha parameter caps the worst-case loss on a market to a known fraction of the liquidity committed.
* Sniper defense: [alpha decay](/how-it-works/alpha-decay) thins the book toward resolution, limiting late-information losses.
## Honest limits
Liquidity is bootstrapped; depth is thin on new markets. The risk is real and bounded, not zero. Mainnet contracts exist; they are not deployed.
# The vault
Source: https://docs.skepsis.live/for-lps/vault-overview
One ERC-4626 vault seeds every market and recycles capital on resolution.
## One vault, every market
Most prediction markets fragment liquidity. Each market has its own pool, and a new market opens dead until someone seeds it. Skepsis pools all liquidity in a single ERC-4626 vault. The vault seeds every market the moment it is created, so a market is tradable on day one.
When a market resolves, its capital returns to the vault and redeploys to live markets. No pool sits idle. No market launches empty.
![LP deposits into one vault, which funds every market and recycles on resolution]()
## What liquidity providers get
* One deposit, exposure across every market, not one.
* Fees earned across all markets the vault backs.
* Capital that recycles on resolution instead of locking up.
## Bounded risk, guaranteed solvency
LMSR keeps the vault's exposure bounded and known. The pool balance is always at least the maximum possible payout, so every market is solvent on every trade. The alpha parameter sets the depth and bounds the worst-case loss on a market to a known fraction of the liquidity committed. See [risks and returns](/for-lps/risks-and-returns) and [economics](/how-it-works/economics).
[Alpha decay](/how-it-works/alpha-decay) adds a second layer: as a market nears resolution its liquidity thins, defending the vault against traders acting on late information.
Liquidity is bootstrapped on testnet. Depth is limited and slippage is real on thin markets. Mainnet contracts exist; they are not deployed.
# Deterministic payouts
Source: https://docs.skepsis.live/for-traders/deterministic-payouts
You see and lock your exact payout the moment you trade.
## You know before you commit
On Skepsis the payout is deterministic. When you take a position, the app shows the exact amount you receive if the outcome lands in your range, and that amount is locked for your trade. No shifting odds are applied retroactively to a position you already hold.
## Why it is deterministic
LMSR fixes your shares at the moment of purchase. Each winning share pays a fixed amount on resolution. Later trades by others move the live quote for the next trade, not the payout on shares you already hold.
## The one thing that moves
The quote moves. The price to open a new position changes as the distribution shifts, so the multiple you can get changes over time. That is why trading early, before the crowd agrees, tends to pay more. See [understanding ranges](/for-traders/understanding-ranges).
# Example: BTC hourly close
Source: https://docs.skepsis.live/for-traders/examples/bitcoin-hour
Price where BTC closes the hour, not just up or down.
BTC is at \$61,000. A binary market asks whether BTC is above \$61,000 at the top of the hour. You think it closes between \$60,800 and \$61,200, a tight band around here.
| | Binary | Skepsis |
| ---------------- | ------------------------- | --------------------------------- |
| Question | Above \$61,000? | Where does BTC close? |
| Your position | Yes | Range: \$60.8k to \$61.2k |
| Outcome \$61,050 | Wins the same as any "up" | Precise. Your tight band pays |
| Outcome \$64,000 | You win | Outside your band. You do not win |
On the binary, a five-cent move over the line pays the same as a \$3,000 run. On Skepsis a 1-of-10 band pays about 10x at a uniform prior, because it is harder to hit. Resolution is the Chainlink BTC/USD feed at the top of the hour.
# Example: CPI print
Source: https://docs.skepsis.live/for-traders/examples/cpi-print
Express a precise CPI view, not just above or below a line.
The next CPI print is expected near 3.0 percent. A binary market asks whether CPI comes in above 3.0 percent. Your read is tighter: 2.8 to 3.2 percent.
| | Binary | Skepsis |
| ------------- | ---------------------------- | --------------------------------- |
| Question | CPI above 3.0%? | Where does CPI print? |
| Your position | No | Range: 2.8% to 3.2% |
| Outcome 2.9% | Wins the same as any "below" | Precise. Your band pays |
| Outcome 1.5% | You win | Outside your band. You do not win |
Binary collapses the whole distribution to one line. A trader who modeled 2.9 percent earns the same as one who guessed "below." Skepsis pays for the precise read. This is the macro use case: express the full view, not a coin flip.
# Example: stETH yield
Source: https://docs.skepsis.live/for-traders/examples/steth-yield
A treasury hedge that binary markets cannot price.
A treasury holds stETH and needs the staking yield to stay in a band to meet its plan. Say it needs yield between 3.0 and 3.4 percent.
A binary market only offers "yield above 3.2 percent, yes or no." If yield lands at 3.1 percent, inside the treasury's acceptable band, the binary may pay nothing. That mismatch between the contract's line and the real exposure is basis risk.
On Skepsis the treasury takes the 3.0 to 3.4 percent range directly. The hedge pays when the yield lands where the exposure actually lives, not when it crosses an unrelated line. This is the hedging use case, and the first step toward parametric insurance.
# Market types
Source: https://docs.skepsis.live/for-traders/market-types
Range, time, and options. Every number you have a view on is tradable.
Skepsis prices three shapes of outcome. All three use the same engine: LMSR over a bucket tree, one continuous curve, one shared pool of liquidity.
Where a number lands. BTC monthly close, CPI print, ETH/BTC ratio, stETH yield. You pick the band.
When a level is reached. The first time a price tags a threshold, settled on a Chainlink feed with a time gate.
One of N named outcomes. All options sit on one curve with shared depth, not N separate yes/no books.
## Why one curve matters
Stack ten binary markets to approximate a range and you fragment the liquidity ten ways. Each book is thin; each spread is wide. Skepsis prices the whole range on a single curve, so depth is shared across every bucket. The same holds for options: N choices, one market, one pool, instead of N separate contracts competing for the same capital.
That is what makes **every range tradable**: you are not waiting for a counterparty on your exact strike. The curve always quotes a price.
## What is not listed
Skepsis is curated and numerical. Markets resolve from a named oracle feed. Subjective, attribution, and sports-outcome markets are out of scope by design.
# Placing a trade
Source: https://docs.skepsis.live/for-traders/placing-a-trade
Choose a range, see the payoff, and take a position.
Every trade is the same four steps, whether the market is a price range, a time, or a set of options.
![Skepsis trading screen: live bucket distribution and the Your Trade panel]()
Pick a market from [alpha.skepsis.live/markets](https://alpha.skepsis.live/markets). Each market shows the current distribution across buckets.
Select the bucket or band you expect the outcome to land in. A tighter range pays more, because it is harder to hit.
Enter the amount of USDC. The app shows your payoff before you confirm: what you receive if the outcome lands in your range.
Confirm the transaction. Your position is issued as an ERC-1155 token and appears in your portfolio.
## What you are paid
If the outcome lands in your range, you are paid in proportion to how tight your range was. At a uniform prior, a 1-bucket pick on a 10-bucket market pays 10x; a 2-bucket pick pays 5x. Payoff = N / k, for k of N buckets.
The price moves as others trade, so the quote you see is the quote at that moment. The payoff shown at confirmation is the payoff you lock in for that trade.
## Selling early
You can close a position before resolution at the current price, the same way you opened it. You do not have to wait for the market to settle.
Liquidity is bootstrapped on testnet. On thin markets, large trades move the price; size accordingly.
# Reading the curve
Source: https://docs.skepsis.live/for-traders/reading-the-curve
How to read the distribution and find a range worth taking.
## The curve is the crowd's view
Each market shows a distribution across buckets: the crowd's current estimate of where the number lands. A tall bucket is a confident consensus. A flat curve is low conviction.
## What to look for
* A peak that looks too narrow: the crowd may be overconfident, and a wider band underpriced.
* A flat region where you have a real read: an underweighted bucket pays more for the same outcome.
* Recent moves: the curve updates as others trade, and a fresh shift can leave a bucket mispriced for a moment.
## From curve to position
Pick the band you believe the outcome lands in, then check the live payoff the app quotes. Tighter band, larger payoff, lower hit probability. Your edge is a range you think is underpriced relative to where it actually lands. See [placing a trade](/for-traders/placing-a-trade).
# Strategies
Source: https://docs.skepsis.live/for-traders/strategies
Common ways to take a position on a continuous market.
These are mechanics, not advice. Markets carry real risk.
## Precision
Take the tightest band you can defend. Highest payoff, lowest hit rate. Best when you have a specific level in mind and the crowd is spread out.
## Spread
Take a wider band across several buckets. Lower payoff, higher hit rate. Best when you have a direction and a rough zone but not an exact level.
## Early conviction
Trade before the crowd converges. Early, the distribution is flat and your band is cheaper, so the same outcome pays more. Alpha decay also makes early trades lower-impact. See [alpha decay](/how-it-works/alpha-decay).
## Close early
You do not have to hold to resolution. If the curve moves your way, close the position at the better price and lock the move. See [placing a trade](/for-traders/placing-a-trade).
# Understanding ranges
Source: https://docs.skepsis.live/for-traders/understanding-ranges
How buckets, the curve, and payoff = N/k fit together.
## Buckets and the curve
A market splits its outcome into buckets of equal width. Together the buckets form a distribution: the market's current view of where the number will land. Buying a bucket raises its share of the curve and its price; the rest adjust so the curve always sums to one.
You take a position on one bucket or a band of adjacent buckets. The band is your range.
![Distribution curve with a highlighted range and its payoff]()
## Payoff scales with precision
The tighter your range, the larger the payoff, because fewer outcomes pay it. At a uniform prior:
* 1 bucket of 10 pays 10x.
* 2 buckets of 10 pay 5x.
* 5 buckets of 10 pay 2x.
Payoff = N / k, for k of N buckets, at uniform prior. As others trade, the prior shifts away from uniform and the live payoff moves with it. The app always shows the live number before you confirm.
## Reading the distribution
![Live bucket distribution with a selected range and implied probability]()
A tall bucket is where the market expects the outcome. A flat distribution means low conviction; a peaked one means the crowd agrees. Your edge is a range you believe is underpriced relative to where it will actually land. See [reading the curve](/for-traders/reading-the-curve).
# Alpha decay
Source: https://docs.skepsis.live/how-it-works/alpha-decay
Liquidity thins toward resolution to defend LPs against late-information snipers.
Alpha is the LMSR liquidity parameter. Higher alpha means deeper liquidity and smaller price impact per trade. Skepsis decays alpha over a market's life.
## How it works
Alpha decays linearly from an initial value to a floor over a configured window. The floor is at least ten percent of the initial value, so the market never goes fully illiquid. Decay is applied in epochs and checkpointed on-chain.
```
Start of life: alpha = high (deep book, small price impact)
Toward end: alpha = floor (shallow book, large price impact)
```
## Why it decays down
Near resolution, information asymmetry is at its worst. A trader with a late read on the outcome can pick off liquidity providers, and a deep book would let them size up. Thinning the book as resolution approaches limits how much a late-information trader can extract. The contract calls this sniper defense.
The cost lands where it belongs. Early, when information is even, the book is deep and trading is cheap. Late, when information is lopsided, the book is shallow and large trades pay for the asymmetry they carry.
## What it means for you
* Trading early is cheaper: deeper book, smaller impact.
* Trading late costs more price impact, by design.
* Liquidity providers hold a bounded, defended position through resolution. See [the vault](/for-lps/vault-overview).
# Continuous vs binary
Source: https://docs.skepsis.live/how-it-works/continuous-vs-binary
The same outcome, priced two ways. One carries basis risk; one removes it.
## Binary: one line
A binary market splits a numerical outcome at a single strike: above or below. It prices direction. It cannot say how much, and it pays the same whether the number lands a tick past the strike or far beyond it.
## Continuous: the whole curve
Skepsis prices the entire distribution as buckets on one curve. It prices magnitude. Payoff scales with precision: a tighter range is harder to hit, so it pays more. At a uniform prior, payoff = N / k for k of N buckets.
![Binary step payoff versus Skepsis continuous payoff]()
## The difference that matters
The gap between the outcome you priced and the outcome that paid is basis risk. Binary carries it. Continuous removes it. For a trader, that is the difference between being paid for direction and being paid for accuracy. For a hedger, it is the difference between a contract that tracks the exposure and one that does not.
# Economics
Source: https://docs.skepsis.live/how-it-works/economics
Where payouts come from, how the vault earns, and why it stays solvent.
## Where payouts come from
When you trade, the LMSR market takes the other side and collects your USDC into the market's pool. The pool is seeded by the vault and held at or above the maximum possible payout at all times, so the market can always pay every winning share. That is the solvency guarantee.
On resolution, the winning bucket is set from the Chainlink feed. Holders of a winning bucket claim a fixed amount per share. Collateral reserved for outcomes that can no longer happen is released as surplus.
## How the vault earns
The vault seeds every market and, on resolution, harvests the surplus and the market's residual capital back for reuse. Across a market, the vault's net is the spread the LMSR captured on trading, plus fees. Because one vault backs every market and recycles on resolution, the same capital earns across many markets instead of locking into one. See [the vault](/for-lps/vault-overview).
## Why it stays solvent
LMSR keeps the pool at or above the maximum possible payout at every point. A market cannot promise more than it can pay. The alpha parameter sets the depth and bounds the worst-case loss on a market to a known fraction of the liquidity committed. [Alpha decay](/how-it-works/alpha-decay) thins the book toward resolution, limiting late-information losses.
## The honest limits
* Liquidity is bootstrapped. Depth is limited on new and thin markets.
* The vault takes real, bounded risk. Fees and spread are the compensation for it.
* Mainnet contracts exist. They are not deployed.
# LMSR for humans
Source: https://docs.skepsis.live/how-it-works/lmsr-for-humans
The pricing rule behind the curve, without the math degree.
## The job of the pricer
A market needs a price for every bucket, at every moment, that always forms a coherent distribution. LMSR, the Logarithmic Market Scoring Rule, does this. It reads the shares held in each bucket and returns a price for each, on one continuous curve.
## What you feel as a trader
* Buying a bucket raises its price; selling lowers it. The curve rebalances so probabilities sum to one.
* Bigger trades move the price more. Depth is set by alpha, the liquidity parameter.
* There is always a quote. You never wait for a matching counterparty; the rule prices both sides.
## Why LMSR, not an order book
An order book over a numerical range needs a resting bid and ask on every bucket. Thin markets leave gaps. LMSR is a single automated maker over the whole range, so depth is shared across every bucket and every range stays tradable.
## Alpha sets the depth
High alpha means deep liquidity and small price impact. Low alpha means a shallower book and larger impact. Alpha is configured per market and decays over the market's life. See [alpha decay](/how-it-works/alpha-decay).
# Trust and resolution
Source: https://docs.skepsis.live/how-it-works/trust-and-resolution
How markets settle, and why Skepsis is curated.
## Resolution
Every market resolves from a named Chainlink feed at a set time. The outcome is numerical and the resolution time is fixed when the market is created, so settlement is mechanical: the feed reports, the contract settles, positions pay out.
No discretionary judging. No dispute desk deciding what happened. The oracle and the clock decide.
## Curated, not permissionless
Not anyone can list a market. Every market on Skepsis is reviewed and listed by the team, against a named feed and a clear resolution rule. Permissionless listing is the largest unbounded-risk surface in this category. Curation removes it, and the customers who hedge real exposure want it as a trust signal.
## What you can verify
* The resolving feed and the resolution time are set on-chain at creation.
* Pricing is LMSR over a bucket tree; the pool is always at least the maximum payout.
* Positions are ERC-1155 tokens; the vault is an ERC-4626 contract on Arbitrum Sepolia.
## Honest limit
Resolution waits on the oracle's time gate. A market settles when its feed reports at the scheduled time, not before.
# Introduction
Source: https://docs.skepsis.live/introduction
Skepsis is a continuous-outcome prediction market. Price the range a number lands in, not a yes or no.
## What Skepsis is
Skepsis is a continuous-outcome prediction market. Instead of yes or no, you price the range a number lands in. A tighter range pays more, because it is harder to hit.
Binary markets price direction. They pay nothing when the number lands a tick off the strike. Skepsis prices magnitude, so the payoff scales with how close you are. That gap, between the outcome you priced and the outcome that paid, is basis risk. Skepsis closes it.
> Price the number, not the direction.
## One worked example
Bitcoin is at \$95,000. You expect it to close the month between \$98,000 and \$99,000.
| | Binary market | Skepsis |
| --------------- | --------------------------- | --------------------------- |
| The question | Will BTC close above \$97k? | Where does BTC close? |
| Your position | Yes | Range: \$98k to \$99k |
| Outcome \$120k | You win the same | You do not win |
| Outcome \$98.5k | You win | You win, and precision paid |
On the binary, "stops at \$98.5k" is priced the same as "runs to \$120k". Skepsis pays for being right about how much.
At a uniform prior, a band covering a tenth of the likely range pays about ten times your stake, because it is harder to hit. Precision is the edge.
![Binary step payoff versus Skepsis continuous payoff]()
## Two ideas to take away
Range, time, and one-of-N options markets, priced on one continuous curve. Any number you have a view on.
A single ERC-4626 vault seeds every market and recycles capital on resolution. No dead pools.
## The essentials
* Chain: Arbitrum Sepolia. Resolution: Chainlink, numerical and time-gated.
* Pricing: LMSR over a bucket tree. Positions: ERC-1155. Liquidity: one ERC-4626 vault.
* A curated set of live markets across crypto and macro. Numerical outcomes only.
* Private beta. Access opens in waves.
Claim test USDC and take your first position. About a minute.
# Glossary
Source: https://docs.skepsis.live/overview/glossary
The terms used across Skepsis, defined briefly.
**Continuous-outcome market.** A market that prices the range a number lands in, not a yes or no.
**Range, time, options.** The three market shapes. Range: where a number lands. Time: when a level is reached. Options: one of N named outcomes, priced on one shared curve.
**Bucket.** One interval of the outcome range. Markets are split into buckets; you take a position on one or more.
**LMSR.** Logarithmic Market Scoring Rule. The pricing rule that turns shares held in each bucket into a price, on one continuous curve.
**Alpha.** The LMSR liquidity parameter. Higher alpha means deeper liquidity and smaller price impact per trade.
**Alpha decay.** A scheduled reduction of alpha over a market's life, from an initial value to a floor. It thins the book near resolution to defend liquidity providers against late-information snipers.
**Basis risk.** The gap between the outcome you priced and the outcome that paid. Binary markets carry it; Skepsis is built to remove it.
**Uniform prior.** The starting state where every bucket is equally likely. Payoff math is quoted at uniform prior: payoff = N / k for k of N buckets.
**Vault.** A single ERC-4626 contract that seeds every market and recycles capital on resolution.
**Position.** Your holding in a market, issued as an ERC-1155 token.
**Resolution.** Settlement of a market from a Chainlink numerical feed at a set time.
**Curated.** Every market is reviewed and listed by the team. Skepsis is not permissionless.
# Quick start
Source: https://docs.skepsis.live/overview/quick-start
Claim test USDC and take your first position on Arbitrum Sepolia.
Skepsis runs on Arbitrum Sepolia. Trading uses test USDC, so you can learn the mechanism with nothing at risk.
Go to [alpha.skepsis.live/markets](https://alpha.skepsis.live/markets).
Sign in with email or Google. A wallet is created for you, or connect your own.
Claim test USDC from the faucet in the app.
Pick a market that fits your view, a crypto price, a macro print, or a yield. Choose the range you expect, set your size, and confirm. Your payoff is shown before you commit.
## Access
Skepsis is in private beta. Access opens in waves. If you have a code, enter it in the app. If not, request access at [skepsis.live](https://skepsis.live).
Liquidity is bootstrapped on testnet. Slippage is real on thin markets.
# Why continuous markets
Source: https://docs.skepsis.live/overview/why-continuous-markets
Binary markets price direction. Skepsis prices magnitude. The difference is basis risk.
## The problem with yes or no
A numerical outcome has a continuum of resolutions. A binary market collapses it to one line: above or below a strike. A tick past the strike or a mile past it, the payoff is identical.
The trader who modeled the exact level earns the same as the one who guessed the direction. The hedger who needed the number to land in a band gets paid only if it crosses a line that may have nothing to do with the exposure.
That gap, between the outcome you priced and the outcome that paid, is **basis risk**.
![Binary step payoff versus Skepsis continuous payoff]()
## How Skepsis prices it
Skepsis splits the outcome into buckets and prices the whole distribution as one curve. You pick the range you expect. Payoff scales with precision: a tighter range is harder to hit, so it pays more.
At a uniform prior on a 10-bucket market, a 1-bucket pick pays 10x. A 2-bucket pick pays 5x. A 5-bucket pick pays 2x. Payoff = N / k, for k of N buckets.
## Who this serves
* Traders expressing a precise view. CPI in 2.8 to 3.2 percent, not CPI above 3 percent.
* Treasuries and LST protocols hedging a precise exposure, without binary's basis risk.
* In time, parametric insurers transferring catastrophe risk priced as a curve, not a coin flip.