Recommended Free Tools
iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
A moving reference is an input to a market maker’s estimate of a contract’s fair value—not the price of a Polymarket outcome token, and not necessarily the price used to settle that market. To quote around it, define the reference and its timing, translate it into an outcome probability, then account for uncertainty, inventory, executable order-book depth and data freshness. Treat any such method as your own pricing framework: Polymarket’s documentation describes its CLOB and data interfaces, but does not prescribe a universal TWAP quoting formula.
What does “TWAP” mean in a Polymarket quoting setup?
TWAP can refer to two different things. A time-weighted average price reference is a measure of an underlying asset or other input calculated over a defined window. A time-weighted average price execution schedule divides an order into smaller child orders spread over time. The first may inform a market maker’s estimate of an outcome’s probability; the second is an execution technique. Neither term, by itself, tells you how a particular Polymarket contract resolves.
Keep three prices separate:
- Reference price: the underlying observation or calculated value you use as an input—for example, a spot price, index, oracle value or windowed average.
- Outcome-token price: the price at which a Yes or No contract token is quoted or traded in Polymarket’s binary order book. It reflects the market’s price for that contract, not the underlying asset’s spot price.
- Execution schedule: the timing and sizing of orders used to work an order over time. A schedule does not define either the market’s settlement rule or a fair-value model.
Do not call a market a “TWAP market” simply because your model uses a TWAP reference. Before trading, read that market’s resolution rules and identify the stated data source, observation times and averaging method, if any.
Free tools Windows power users keep installed
One-click scans. No signup required.
How does Polymarket’s order book affect the quote?
Polymarket describes its exchange as a hybrid-decentralized central limit order book. The operator handles order matching and ordering off-chain; execution and settlement take place on-chain under users’ signed limit-order instructions. Its documentation says the operator cannot choose a user’s price or execute outside those signed instructions, and describes matching complementary outcome tokens.
#1 Best Overall
- Language: english
- Book - trading: technical analysis masterclass: master the financial markets
- It is made up of premium quality material.
For pricing purposes, that means your estimated probability is not itself an order-book quote. You must express a price in the relevant outcome token and make sure your order complies with the applicable market rules and tick size. The CLOB’s displayed prices and available size are also distinct from the underlying reference value you use in a model.
How should you construct a quote around a moving reference?
The following is an analytical workflow, not an official Polymarket formula or a strategy shown to be profitable. Its purpose is to make assumptions and operational risks explicit.
Rank #2
- Specify the reference. Record the instrument, source, timestamp convention, update cadence and whether the input is spot, an index, an oracle or a windowed average. If it is a TWAP, define the start and end of the averaging window, observation frequency and treatment of missing or stale readings. Confirm that the contract’s resolution rules actually use the feed or calculation you intend to model.
- Translate reference changes into contract probability. Estimate the probability of each outcome conditional on the reference and time remaining. A move in the underlying price does not translate one-for-one into a move in binary probability: the result depends on the event definition, resolution condition, volatility assumptions and time to resolution. Write down those assumptions rather than treating a reference-price change as a token-price change.
- Set a fair-value center, then account for risk. Let your model’s estimated Yes probability be
p. As a conceptual starting point, a Yes quote can be centered aroundpand a No quote around the complementary probability, subject to the contract’s rules and the actual book. Widen or skew those quotes as appropriate for reference uncertainty, data latency, inventory, adverse selection and expected execution costs. The framework does not determine a particular spread, hedge ratio or latency threshold. - Check both outcome books and executable depth. Use current token-specific best prices and available size for Yes and No; inspect the applicable tick size, fees and any market-specific incentives. Compare executable levels rather than relying on an indicated midpoint or a historical price point. Account for the complementary relationship between the outcomes, but do not assume that displayed bids, asks, fees and available depth make both sides immediately tradable at perfectly complementary prices.
- Define refresh and failure controls. Decide which changes in the reference, order book, inventory, market status or data freshness trigger a cancel-and-replace. Set quote-age and size limits, and specify a fail-closed action—such as withdrawing affected quotes—if your reference feed or order-book stream becomes stale. These are controls for your own system, not documented Polymarket behavior.
- Evaluate execution, not just forecast accuracy. Track fill probability, realized spread, markout after fills, inventory drift and execution shortfall separately. A backtest should account for queue position, partial fills, fees and timestamp alignment; a price series alone cannot establish that a quote could have traded at the recorded level.
How can you retrieve the relevant Polymarket data?
Polymarket Institute’s guide to exchange research data describes CLOB price requests as token-specific: a request uses the ID for the Yes or No outcome. It identifies Gamma’s clobTokenIds field as a way to find those token IDs, demonstrates a best-price request using /price, and shows historical-price queries using /prices-history. The guide also points to the Data API for trade-history and user-history data.
Use the official documentation for current request details and the live platform for current values; examples in a guide are not current market quotes. The Institute’s guide points readers to Polymarket’s Orderbook & Pricing documentation for fees, tick sizes and spreads. Polymarket’s Trading help collection links to information on limit orders, liquidity rewards, maker rebates and trading fees, but the collection page alone does not establish current terms for a particular market or account. Verify applicable details before using them in pricing or backtests.
Rank #3
- Used Book in Good Condition
A best-price response is not a substitute for depth when you need to trade size. For an execution-aware quote, inspect available levels on both outcome tokens and check that your data timestamps align with the reference observations and your own order events.
Fixed-spread or reference-adjusted quoting: which fits?
These are two broad approaches, not mutually exclusive recipes. A fixed-spread quote keeps a chosen distance around its pricing center; a reference-adjusted quote updates that center as its reference-based fair value changes. Either can still use inventory skew and risk limits. The trade-offs depend on market conditions and implementation, and the cited sources establish no universally optimal setting.
Rank #4
| Consideration | Fixed-spread quote | Reference-adjusted quote |
|---|---|---|
| Response to genuine fair-value changes | Does not move its center just because the reference changes; may require separate manual or rule-based updates. | Can update its center as the modeled probability changes, if the reference-to-probability mapping is sound. |
| Stale-reference adverse selection | Can be exposed if the reference moves and the quote is not otherwise refreshed. | Can also be exposed when its feed or model lags; automatic adjustment is not protection against stale inputs. |
| Inventory sensitivity | Can add inventory skew, but a fixed center does not inherently account for changing fair value. | Can combine reference-based repricing with inventory skew; those inputs need separate controls. |
| Quote stability and noise | May avoid reacting to every reference fluctuation, although book and risk changes can still prompt updates. | May change more often if its reference is noisy or updates frequently; filtering introduces its own lag and assumptions. |
| Execution probability and queue position | Stability may reduce unnecessary repricing, but competitiveness still depends on other quotes and available size. | More frequent cancel-and-replace can affect queue position; a fair-value update does not guarantee a fill. |
| Fees and incentives | Must account for applicable fees and any verified incentives. | Must account for the same market-specific terms; repricing alone does not establish an incentive or net return. |
| Operational complexity | Requires quote controls and monitoring, but may use fewer reference-driven updates. | Also requires feed validation, timestamp handling, model updates and safeguards against stale or noisy inputs. |
Should the reference be deterministic or randomized?
If you calculate a reference from scheduled observations, window length, sampling cadence and timestamp alignment affect how quickly it reflects new information and how sensitive it is to noise. A longer averaging window generally incorporates observations over more time than a shorter one, but no source cited here identifies an optimal window for Polymarket market making. The right specification depends first on the contract rules and then on the model’s intended use.
Do not conflate randomizing execution times with randomizing the reference itself. The Bank for International Settlements Markets Committee report FX execution algorithms and market functioning (30 October 2020) discusses execution algorithms in FX markets, not Polymarket. It says TWAP slicing is intended to reduce market impact, while an overly aggressive schedule can still have substantial impact. It also notes that randomizing execution timing can make an order less predictable. As the report puts it: “Current versions of TWAP algorithms typically provide for some randomisation in the timing of execution (ie will not attempt to execute exactly every 60 seconds in this case), to reduce the predictability and signalling from orders.”
Best Value
- Prentice Hall Press
- Ideal for a bookworm
- It's a great choice for a book person
That FX observation is a useful execution analogy, not evidence that randomized timing improves Polymarket results. A deterministic or sampled reference input is a separate design decision: compare its window, observation cadence, lag, noise sensitivity and predictability for the specific market and data stream, and do not treat any setting as proven optimal.
What should you measure before changing the quoting method?
Judge the system on realized execution and risk, not only on whether its reference estimate looked plausible after the fact. Use timestamped records of reference observations, model values, book states, quote submissions and cancellations, fills, fees and inventory. Keep the measures distinct so a high fill rate does not conceal poor post-fill markouts or costly inventory accumulation.
- Fill probability: how often and how much of a quote executes under the conditions you are studying.
- Realized spread and markout: execution economics at the fill and subsequent price movement, measured over explicitly chosen horizons.
- Inventory drift: how fills change exposure and whether that exposure persists or grows.
- Execution shortfall: the difference between the execution achieved and a clearly defined benchmark, with fees and timing handled consistently.
Polymarket’s published material cited here does not provide a topic-specific statistic establishing the performance of TWAP-based market making on its exchange. The BIS report’s broader observations about algorithmic execution and the transfer of execution risk concern FX markets; they should not be presented as measured Polymarket findings.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchQuick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
