Backtest a trader - 0xinsider API

By the end you have a wallet’s daily profit and loss curve, the large trades 0xinsider stored for that wallet, and its fill-by-fill position in any one market where the two do not line up. Run it before you follow a wallet, or when you want to check the grade it was given. You need a Pro API key. See Authentication.

1. Read the profit and loss series

Trader P&L returns a precomputed daily series for one wallet’s Polymarket trading. A wallet 0xinsider knows but has no series for returns 200 with an empty structure. An address that does not resolve to any wallet returns 404.

Field	Meaning
`entries`	One point per day, oldest first. `daily_change` is always a number. `cumulative_profit` can be absent.
`stats`	Four windows: `all`, `d90`, `d30`, and `d7`. Each carries `current`, `change`, `best_day`, `worst_day`, `win_days`, and `loss_days`, among others.
`drawdown`	How far below its own peak the wallet was on each day. `monthly` and `year_totals` sum the same series by month and by year.
`freshness_at`	When 0xinsider last rebuilt this series. It is always present, and `null` until the first rebuild.

import os

import requests

BASE = "https://api.0xinsider.com/api/v1"
H = {"Authorization": f"Bearer {os.environ['OXINSIDER_API_KEY']}"}

addr = "0x204f72f35326db932158cba6adff0b9a1da95e14"
pnl = requests.get(f"{BASE}/trader/{addr}/pnl", headers=H, timeout=30).json()["data"]

print("as of:", pnl["freshness_at"])
print("all-time:", pnl["stats"]["all"]["current"])
print("last 30 days:", pnl["stats"]["d30"]["change"])

for entry in pnl["entries"][-10:]:
    # cumulative_profit is left out of the JSON on a day with no value for it.
    print(entry["date"], entry.get("cumulative_profit"), entry["daily_change"])

for point in pnl["drawdown"][-3:]:
    print(point["date"], point["cumulative_profit"], point["drawdown"])

2. Read the stored large trades

Whale trades history returns the large trades 0xinsider stored for one wallet over a window you choose, newest first.

Parameter	Description
`trader`	A wallet address or a username. A wallet 0xinsider has never seen returns an empty list rather than an error.
`from`, `to`	RFC3339 timestamps. `from` is included, `to` is excluded, and `to` must be later than `from`.
`min_size`	The smallest trade to return, in USD. The default is $5,000.
`platform`	`polymarket` or `all`. They return the same rows, because Polymarket is the only provider. Any other value returns `400`.
`limit`	Trades per page, from 1 to 100. The default is 20.

The response’s meta.completeness.status is always best_effort. How much of a wallet’s history is stored varies by market and by window, so treat a gap as missing data rather than as a period when the wallet did not trade.

fills = []
cursor = None
while True:
    params = {
        "trader": addr,
        "from": "2026-01-01T00:00:00Z",
        "to": "2026-06-01T00:00:00Z",
        "limit": 100,
    }
    if cursor:
        params["cursor"] = cursor
    page = requests.get(
        f"{BASE}/whale-trades/history", params=params, headers=H, timeout=30
    ).json()
    fills += page["data"]
    if not page["has_more"]:
        break
    cursor = page["next_cursor"]

for tr in fills:
    print(
        tr["traded_at"], tr["side"], tr["outcome"], tr["size_usd"],
        "@", tr["price"], tr["market"]["title"],
    )

3. Rebuild one market position

Position timeline returns every stored fill for one wallet in one market, newest first. The condition_id parameter is required.

Field	Meaning
`running_amount`	The wallet’s share balance after this fill, counting fills from earlier pages too. It leaves out splits, merges, redemptions, and neg-risk conversions, so it is not the wallet’s complete holding.
`running_avg_price`	The buy-weighted average price of the stored fills. A sell does not change it, and it is not the current average price Polymarket shows for the position.
`amount_delta`	The change in shares, positive on a buy and negative on a sell. `usdc_notional` is the value of the fill, always positive.

curl -H "Authorization: Bearer $OXINSIDER_API_KEY" \
  "https://api.0xinsider.com/api/v1/trader/0x204f72f35326db932158cba6adff0b9a1da95e14/position-timeline?condition_id=0x123...&limit=100"

0xinsider keeps a timeline only for the wallets it syncs most closely. For any other wallet the route answers 404 with error.reason trader_not_tracked. Do not retry that wallet: the answer will not change, and no other spelling of the address reaches a timeline that was never recorded.

4. Compare the two views

Group the trades from step 2 by the date part of traded_at.
Match each date to the entries point with the same date.
For any day with a large daily_change, open the timeline in step 3 for each market.condition_id you saw that day.

A daily_change with no matching trade is normal, and it does not mean the numbers disagree. Step 1 counts every position the wallet holds, while step 2 lists only the large trades that were stored.

What this does not do

Tell you what following the wallet would have paid. You get what the wallet did, at its own fill prices, from a stored history marked best_effort. That is a record, not a simulation of your own returns.
List a wallet’s closed positions. Read them one market at a time with the timeline in step 3.
Return the rows of a full export. Trader export snapshot describes what an export would cover, and Submit a trader export queues the job that produces the dataset.
Carry a token_id on every timeline event. It is null when 0xinsider has no token id stored for that market.
Place an order. No endpoint on this API does.

Copy-trade signals builds a follow list from the leaderboard. Grades explains the letter on each wallet.

​1. Read the profit and loss series

​2. Read the stored large trades

​3. Rebuild one market position

​4. Compare the two views

​What this does not do

​Related

1. Read the profit and loss series

2. Read the stored large trades

3. Rebuild one market position

4. Compare the two views

What this does not do

Related