/prices-history end point returns empty data for resolved markets at granularities below 12 hours

[MuhammadOmarMuhdhar]

Problem

The /prices-history endpoint only returns historical data at 12+ hour granularity for resolved/closed markets, even for extremely high-volume events where price fluctuations definitely occurred at finer intervals.

Example:

Market: 2024 US Presidential Election (https://polymarket.com/event/presidential-election-popular-vote-winner-2024)
Token ID: 21271000291843361249209065706097167029083067325856089903026951915683588703117 (yes on will Kamala Harris win popular vote)
Status: Resolved/Closed

Despite this market having massive trading activity with prices fluctuating minute-by-minute during the election, I cannot retrieve historical data at hourly (or any sub-12-hour) granularity.

Code:

import requests

url = "https://clob.polymarket.com/prices-history"
token_id = '21271000291843361249209065706097167029083067325856089903026951915683588703117'

# This works - returns data
params_12h = {
    "market": token_id,
    "interval": "max",
    "fidelity": 60*12  # 12 hours in minutes
}
response = requests.get(url, params=params_12h)
print("12-hour:", response.json())
# Returns: {'history': [{'t': ..., 'p': ...}, ...]} 

# This fails - returns empty
params_1h = {
    "market": token_id,
    "interval": "max",
    "fidelity": 60  # 1 hour in minutes
}
response = requests.get(url, params=params_1h)
print("1-hour:", response.json())
# Returns: {'history': []} 

Is this a data retention policy issue, or is there a way to access finer granularity for closed markets?

Thanks!

[MuhammadOmarMuhdhar]

Figured it out!

Solution: Use startTs and endTs parameters instead of interval:

import requests
from datetime import datetime, timedelta

url = "https://clob.polymarket.com/prices-history"
token_id = "YOUR_TOKEN_ID"

# Specify exact time range instead of "max"
start_time = int(datetime(2024, 11, 1).timestamp())
end_time = int(datetime(2024, 11, 14).timestamp())

params = {
    "market": token_id,
    "startTs": start_time,
    "endTs": end_time,
    "fidelity": 60
}

response = requests.get(url, params=params)
print(response.json())

For longer time ranges, chunk requests:

all_data = []
chunk_days = 15  # Request 15 days at a time

while current_start < end_date:
    current_end = min(current_start + timedelta(days=chunk_days), end_date)
    # Make API request for chunk
    all_data.extend(response['history'])
    current_start = current_end

Note: I'm curious why interval: "max" fails with low fidelity while explicit time ranges works, perhaps there's a data retention policy or query timeout that kicks in when interval: "max" is combined with high-resolution data requests?