Polymarket has no native demo mode, so the only honest way to paper trade the venue is to simulate your own fills against the live order book and journal every entry as if real money were on the line. The methods that work are spreadsheet journaling against current quotes, shadow fill simulation that respects slippage and fees, and a strict log of at least thirty trades before any live capital touches the wallet. Anything less generous to reality is a paper trader who plans to lose money quietly the first month after going live.
Why Polymarket has no native demo mode
Most centralised venues ship a built-in paper trading mode because the matching engine is theirs and the simulation costs them nothing. Polymarket is different. The order book is on-chain-adjacent, settlement is real USDC on Polygon, and there is no parallel testnet that mirrors live liquidity. The venue would have to spin up an entire shadow universe of markets, fills, and resolutions to give traders a demo button, and that universe would never reflect the real one closely enough to be useful. So Polymarket made the pragmatic choice and shipped nothing.
The absence is honest, but it leaves a real gap. New traders who would happily push fake money around a sandbox for two weeks before going live instead either skip practice entirely and donate their first deposit to the spread, or they install a half-broken third-party simulator that ignores fees and prints fantasy P&L. Both paths end the same way. The fix is to build your own paper trading process that respects the parts of the live venue that actually matter, and to be ruthless about what counts as a realistic fill.
Paper trading as a discipline is well documented outside crypto. The Investopedia entry on paper trading is the canonical reference and the framing carries over: a paper trade only teaches you something if it would have actually happened at the price you wrote down. Most home-grown Polymarket simulators violate that rule the moment they assume a midprice fill on a thin market.
What honest paper trading actually requires
Three components separate paper trading that prepares a trader from paper trading that builds bad habits. Skip any of them and the log is fiction.
The first is a real-time price source. You need to be looking at the same order book you would be hitting if the trade were live, at the moment you write the entry. Hours-old screenshots do not count. A market that traded at 42 cents this morning may be 51 cents now because a poll dropped, and a paper log that entered at the morning price is logging a trade that nobody could have taken.
The second is fill realism. Polymarket charges no maker fees and no taker fees on most markets, which sounds free, but slippage on thin books eats five to fifteen cents per share routinely. A naive paper trader writes down the midprice as their entry. A realistic one writes down the price they would have paid after walking the book to fill their stated size. Those two numbers diverge wildly on small markets, and the divergence is exactly where most newcomers blow up live.
The third is a journal that captures enough state to learn from. Entry price alone is not enough. You need timestamp, side, size, the order-book depth at the moment of entry, your stated thesis in one sentence, exit price, exit reason, and net P&L after slippage. The journal is the asset. The trades are the inputs.
Method 1: Spreadsheet journaling against live order book
This is the method I recommend to every new trader who asks how to start. Open a spreadsheet. Open Polymarket in another tab. When you see a market you would trade, write down the timestamp, the market name, the side you would take, the order-book best bid and best ask at that moment, and your intended size. Then walk the visible book and compute the realistic average fill price for that size. Write that down too. State the thesis in fifteen words or less.
Let the market run. When your thesis resolves either by exit, expiry, or stop, fill in the exit row of the same line. Compute P&L on the slippage-adjusted entry, not the midprice. Mark every trade with a one-sentence post-mortem: what was right, what was wrong, what you would do differently.
Do this for thirty trades minimum before evaluating yourself. Fewer than thirty and you cannot tell skill from variance. The setup effort is fifteen minutes and the ongoing cost is the discipline to actually write down the bad trades, not just the good ones. The single most common failure mode is the trader who logs every winner and conveniently forgets to log the loser. Self-deception in a paper log compounds into self-destruction in a live account.
Method 2: Shadow fill simulation
The next level up is a small script that reads the Polymarket order book over the public API, takes your intended trade as input, and computes a fully realistic fill including slippage, depth consumption, and the post-trade book state. You commit the resulting trade to a local journal and the script tracks unrealised P&L against the live midprice every minute. This is what serious traders do when they want to paper-trade a strategy across more markets than a human can watch by hand.
The funnel below is the workflow I have used to walk three different friends from their first paper trade to a live account they did not blow up. Each stage is gated on the previous one passing, not on time.
From hypothesis to live capital — the six-stage paper trading funnel
What shadow fills look like in code
The simulation is not complicated. You pull the order book for the market with a single REST call. You take the side and size you would trade. You walk the book level by level, summing filled size and weighted price until your order is satisfied. The output is a single average fill price and a leftover-size indicator if the book was too thin to fill your order. That number, not the midprice, is what you write in the journal.
The order book on this venue is documented in the order book explainer. Once you can read it, the simulator becomes a two-hundred-line Python script. The math is the part that catches people: a 1000-share order on a market with 200 shares offered at 42 cents and 800 shares offered at 48 cents has an average fill of 46.8 cents, not 45 cents and not 42. The difference between those numbers is the entire reason paper trading exists.
How the four paper trading methods compare
| Method | Realism | Setup effort | Tracks fees | Tracks slippage | Mobile friendly |
|---|---|---|---|---|---|
| Spreadsheet against live book | High if disciplined | 15 minutes | Manual | Manual | Yes |
| Fantasy prediction markets | Low — different liquidity | 5 minutes | No | No | Yes |
| Polymarket testnet (none exists) | Not available | n/a | n/a | n/a | n/a |
| Shadow fill simulation script | Highest | 1–2 days of coding | Yes | Yes, by walking book | No |
| Copy-of-real-fills journaling | Perfect for live data | Hours per week | Yes | Yes | Partial |
The row to notice is the second one. Fantasy prediction markets where you push fake credits around a parallel venue are popular and useless. The liquidity profile, the resolution criteria, and the participant set are all different. A trader who is excellent at fantasy markets has learned nothing about Polymarket beyond the broad concept of buying low and selling high.
The paper-trading journal schema
This is the spreadsheet schema I give to every new trader who asks. Copy it into a Google Sheet or Excel and do not change the columns until you have logged at least thirty trades. Adding columns early is a procrastination tactic that delays the part that actually teaches you, which is filling rows in.
| Column | Type | What goes in it |
|---|---|---|
| timestamp_entry | datetime | The exact moment you would have hit submit |
| market | text | Full market name as it appears on Polymarket |
| side | YES / NO | Which outcome you are buying |
| size_shares | integer | Number of shares you would buy at this size |
| entry_midprice | decimal | Best bid / best ask average at moment of entry |
| entry_fill_price | decimal | Walked-book average fill for stated size |
| entry_slippage_cents | decimal | fill_price minus midprice in cents |
| fee_entry_usdc | decimal | Currently zero on most markets; record anyway |
| thesis | text, 15 words max | Why this trade in one sentence |
| timestamp_exit | datetime | When you closed or the market resolved |
| exit_fill_price | decimal | Walked-book exit, or resolution value |
| fee_exit_usdc | decimal | Record for symmetry |
| pnl_net_usdc | decimal | (exit - entry) * size minus all fees |
| exit_reason | enum | thesis-hit / stop / time / resolution |
| notes | text | One-sentence post-mortem, especially on losers |
The two columns that matter most are entry_slippage_cents and notes. The first one catches the systematic optimism that infects every new paper trader. The second one is the only thing in the spreadsheet that compounds over time, because it is the only column where the trader is actually thinking.
Common ways paper trading lies to you
Even with a disciplined process, paper trading has structural biases that no spreadsheet column can eliminate. Knowing them is half the protection.
The midprice fantasy. The default failure mode. You write down 42 cents as your entry because that is the number on the screen. The real fill at your size was 46. Four cents per share, multiplied by every trade, is enough to flip a paper winner into a live loser. Walk the book or do not paper trade at all.
The selective journal. The trader who logs every trade where the thesis worked and quietly forgets the ones where it did not. Symptom: a paper log with a 78 percent win rate that turns into a 41 percent live win rate the first month. Cure: log the trade the moment you would have hit submit, before you know the outcome. Treat the journal as a contract with future-you.
The infinite patience trap. In a paper log it costs you nothing to leave a position open for three weeks waiting for the thesis to play out. In a live account, the opportunity cost of locked capital and the emotional cost of watching a position drift against you for twenty days are real. Set a default time-stop in the journal and respect it.
The size-up fallacy. Paper logs are usually run at a size that would not move the book. The moment you go live and try to put real capital to work, your own order moves the price against you. A paper trader who logged 500-share fills should not start live with 5000-share orders.
The skipped loser tax. A subtler version of the selective journal. You see a trade you considered, watch it lose, and tell yourself you would not have taken it anyway. Maybe true, maybe not. The honest fix is to log the entry before you know the outcome, every single time. The detailed treatment of this and related biases is in the risk management writeup.
Signals you are ready to go live
Thirty trades is the floor, not the ceiling. The signals below are what I look for before I tell a paper trader to graduate.
- Thirty trades minimum, fifty preferred. Below that, variance dominates and you cannot tell whether you have an edge or got lucky.
- Net positive after realistic slippage. Not on midprice. After walked-book fills. If you are net negative on slippage-adjusted P&L, going live will be worse, not better.
- Win rate and average win size both stable across the last twenty trades. A win rate that is drifting up because your most recent trades happened to be good is not stability, it is hot-streak coincidence.
- You can articulate why your losers lost. The notes column should read like a real diagnosis, not a vague shrug. If every loser is "got unlucky," you have not learned anything.
- You have not skipped logging a trade in the last month. Discipline gaps in paper trading become risk-management gaps in live trading. The skill is the habit.
If any of those five are missing, keep paper trading. The cost of an extra two weeks of paper is zero. The cost of going live early is the entire account.
Sizing the first live trades
When the signals above are all green, the right first live trade is one tenth of the size you were paper-trading. Not the same size. Not half. One tenth. The reason is not capital preservation, although that is a bonus. The reason is that the emotional weight of real money fundamentally changes decision quality, and the first ten live trades are the calibration period where you find out how much it changes for you specifically.
Run the small-size live trades for at least two weeks before scaling up. Keep the journal running. Compare paper-size P&L to live-size P&L on the same setups. If the live numbers are dramatically worse, you have a discipline problem under real money pressure and the fix is to keep sizing small until the gap closes. If they are roughly equivalent, scale up by a factor of two or three and run another two weeks. Geometric growth, not jumps.
For traders who plan to graduate from manual paper trading to copy-trading a leaderboard wallet rather than running their own thesis, the bridge is covered in the copy trading guide. The paper-trading habit transfers cleanly: the journal becomes a record of which source wallets the bot followed, what fills you got, and how the mirror execution compared to a clean paper baseline.
The point of paper trading is not to prove you can win. It is to prove the journal does not lie. A trader who can keep an honest journal for thirty trades has built the habit that protects the live account. A trader who cannot will lose money no matter what their paper numbers say.
About the author
Eli Marsh founded Poly Syncer after a decade running execution systems for a Chicago prop desk and three years trading prediction markets full time from a personal account. He has logged more than 4,000 trades on Polymarket and writes about the parts of the venue that are not visible from the leaderboard. Eli answers reader questions weekly and reviews every post on this site for accuracy.
Frequently asked questions
Does Polymarket have a demo mode or testnet?
No. Polymarket has no built-in paper trading feature and no public testnet that mirrors live liquidity. The only honest way to practice is to simulate fills against the live order book and journal each trade as if real money were committed.
How many paper trades should I log before going live?
Thirty is the floor and fifty is preferred. Below thirty trades the result is dominated by variance and you cannot distinguish skill from luck. The win rate and average win size should also be stable across the last twenty trades before you graduate.
What is the most common mistake new Polymarket paper traders make?
Logging the midprice as their entry fill instead of the realistic walked-book fill. On thin markets the difference is five to fifteen cents per share, which is enough to flip a paper winner into a live loser. Always compute fill price by walking the order book at your stated size.
Can I paper trade Polymarket on my phone?
Yes for spreadsheet journaling against the live order book, which is the recommended starter method. Shadow fill simulation that runs a script against the API is desktop only. Most new traders should stay with the spreadsheet approach until they have logged thirty trades.
Do fantasy prediction market apps count as paper trading Polymarket?
No. The liquidity, resolution criteria, and participant set on fantasy markets are different enough from Polymarket that performance on one does not predict performance on the other. Only paper trading against the live Polymarket order book builds skills that transfer.