SPECIFICATION & METHODOLOGY

Loota Price Verification Methodology

This technical whitepaper outlines the statistical algorithms, data collection criteria, and validation standards that power Loota Canada (loota.ca).

1. The 90-Day Resampled Baseline

Traditional deal platforms rely on raw observation counts or simple arithmetic averages. This creates a severe sampling bias: products are scanned far more frequently during promotional spikes, falsely depressing the computed average price.

Loota implements Daily Resampled Median Pricing:

Baseline(P) = Median({ MinPrice(day_k) | day_k ∈ Past 90 Days, k ≥ 7 distinct days })

Daily Deduplication: Multiple price observations on the same calendar day are collapsed into a single representative minimum observation.
Minimum Data Threshold: Baselines require at least 5 verified observations spanning at least 7 distinct days. Items failing this threshold are flagged as Insufficient Historical Data rather than generating unsubstantiated ATL claims.
Robust Median: Unlike the arithmetic mean, the median is immune to artificial one-day price spikes intended to game discount calculations.

2. Deal Quality Tiers

Tier	Discount vs 90d Baseline	Historical Requirement	Classification Criteria
S-Tier	≥ 30% below baseline	All-Time Low (ATL)	True historical rock-bottom price on verified genuine hardware.
A-Tier	≥ 20% below baseline	Near 90d Floor	Excellent promotional value, consistently under market median.
B-Tier	10% to 19% below baseline	Below Median	Moderate discount, standard seasonal promotion.
Flagged	> 0% above baseline	Any	Marked as above baseline price. Caution advised.

3. Deceptive Markup & Fake Discount Screening

Under Subsection 74.01(1) of the Canadian Competition Act, ordinary selling price (OSP) representations are deceptive unless the product was offered at the reference price in good faith for a substantial period of time.

Detection Criteria:

Fake Markup (fake_markup): Strikethrough list price is > 25% higher than the 90-day observed ceiling price, creating an artificial discount illusion.
Above Baseline (above_baseline): The merchant's "sale" price is actually higher than the product's regular 90-day median price.
Statistical Anomaly (suspect_price): When Double Median Absolute Deviation (Double MAD) flags an extreme cliff drop without multiple retailer corroborations, the item enters a human verification queue to guard against pricing errors or knockoff listings.

4. Anti-Scam Multi-Evidence Rule

When a deal indicates a discount deeper than 75% below baseline (≤ 25% of 90-day median), Loota requires multi-evidence corroboration before publishing to public feeds:

Verified fulfillment by an authorized retail division or first-party merchant (e.g. Amazon.ca, Best Buy).
Active stock status with standard Canadian domestic delivery.
Title keyword verification confirming the listing is not an accessory, protective case, or knockoff replacement.