Polymarket paper test vs live sim vs Python backtest

Q: Where can I see a failed paper test?

Here: public kill example (https://oddsage.com/strategy/259bc179-04dc-4e6d-816d-59d8f3cb52cc).

Q: Where else is this product described?

Polymarket strategy tester (https://oddsage.com/polymarket-strategy-tester) Test a Polymarket strategy (https://oddsage.com/test-polymarket-strategy) Polymarket paper trading (idea test) (https://oddsage.com/polymarket-paper-trading) OddSage home (https://oddsage.com/)

Three tools can all sound like “paper trading” or “backtesting” on Polymarket. They are not the same job.

This page is an honest split: OddSage (idea keep/kill on resolved history), live virtual-wallet sims (PaperPoly-style), and Python / L2 backtest SDKs. No ranking. No invented features. Hard limits called out where they matter.

Tagline for OddSage stays the same: test your Polymarket idea before you risk real money.

The three jobs in one table

	OddSage	Live virtual-wallet sim	Python / L2 backtest SDK
Job	Paper-test an idea; get keep / kill / no edge / inconclusive	Practice clicking on live odds with fake cash	Replay history in code with your own models
Input	Polymarket URL → Market Dossier, or English idea → draft → Save & run	Deposit play money; pick live markets	Scripts, SDKs, often L2 / order-book data
Output	Readable verdict on resolved history	P&L on markets that have not resolved yet	Charts, fills, metrics you define
Wallet	None	Fake wallet (not OddSage)	None (your infra)
Live trading	No	No real money, but live book practice	No (unless you bolt on execution yourself)
Copy-trade execution	No	Usually no (separate product class)	Only if you build it
Who it is for	Traders with a sentence who want a stop or a hold	Traders learning the UI and timing	Quants who already write Python

OddSage is paper-only Polymarket research. It does not connect a wallet, place live orders, or copy another trader’s book onto yours.

What OddSage actually does

Two doors. Same lab.

Paste a Polymarket URL. You get a Market Dossier: the question, rules, quotes, history. Research page, not a ticket.

Or type the idea in English. Read the draft. Save & run. You get one of four endings: keep, kill, no edge, or inconclusive.

That loop is the product: English idea → draft → Save & run → verdict.

Keep means the paper test left something worth holding. Kill means it did not survive. No edge means it ran and still looked like noise. Inconclusive means there was not enough evidence. Inconclusive is a stop, not a soft keep.

A public example that did not get a pass: this kill / no-edge style paper-test.

Same product, other doors:

What a live virtual-wallet sim does (PaperPoly-style)

Live sims give you fake cash on live odds. You click as if you were in the market. The market has not resolved yet. You are practicing hands, timing, and the UI — not asking whether the idea ever showed an edge on resolved history.

That is a real product. It is not OddSage. OddSage does not give you a play-money bankroll. There is nothing to deposit, even in fake cash. There is a draft, a run, and a verdict.

If you typed “paper trading” and wanted a fake wallet on live odds, use a simulator. If you wanted the idea to be able to fail on history before you fund it, stay on OddSage.

What a Python / L2 backtest SDK does

Python and L2 backtest stacks (Marketlens-style SDKs, Nautilus integrations, notebook repos) let you replay Polymarket history in code. You own the model, the fee assumptions, the fill logic, and the charts. That is the right path if your edge already lives in a repo.

OddSage does not replace that stack. You do not need Python to use OddSage. You also do not get tick-level order-book replay, custom fee models you wrote yourself, or a notebook export of every fill. If that is your workflow, keep it. This page is for people who have a sentence and want keep, kill, no edge, or inconclusive without opening a notebook.

Hard limits (read these)

OddSage hard limits

No wallet connect for trading.
No live order placement.
No copy-trade execution.
No live virtual bankroll.
A keep is still paper research, not a signal to size up.
Inconclusive is a finished result; do not rerun the same draft hoping it blinks keep.
Public kill pages and dossiers only show what the run produced. Missing data stays missing.

Live sim hard limits (typical of the class)

Practice on unresolved markets does not prove historical edge.
Fake P&L is not a keep/kill on resolved history.
A sim that never tells you to stop is practice, not research.

Python / L2 SDK hard limits (typical of the class)

You bring engineering time, data quality, and your own mistakes.
A notebook win rate is only as honest as your fill and fee assumptions.
Not a one-sentence path for a non-coder who just wants a verdict.

Which one should you open?

You have an English idea and want it able to die on paper → OddSage.
You want to practice clicking on live odds with fake cash → a live virtual-wallet sim (not OddSage).
You already backtest in Python and need L2 replay → an SDK / notebook stack (not OddSage).

OddSage will kill an idea. That is the point. A tool that only ever says “interesting” is a mood board.

I’m Avinash Jetwani (@jetwaniavinash). If it dies, it dies on paper.

Create a free account to save a draft and run it.

FAQ

Is OddSage the same as a Polymarket paper-trading simulator?

No. Live sims use a fake wallet on live odds. OddSage paper-tests an idea on resolved history and returns keep, kill, no edge, or inconclusive. No wallet.

Can OddSage place trades or copy another trader?

No. Paper-only research. No live trading. No wallet. No copy-trade execution.

Do I need Python to backtest on OddSage?

No. Type the idea in English, read the draft, Save & run. Python/L2 SDKs are a different path for people who already write code.

What does inconclusive mean?

The run finished. There was not enough evidence to keep or kill. Change the idea, or stop. Do not treat inconclusive as a soft keep.

Where can I see a failed paper test?

Here: public kill example.