The numbers behind the team

93,728+ MLS transactions analyzed. 49 peer-reviewed research papers. 1,000+ internal team closings. Every stat traces back to those three foundations.

View:
Foundation A · 93,728+ MARIS transactions Foundation B · 49 peer-reviewed research papers Foundation C · 1,000+ internal Artemis closings
Top Performance Stats

Team scoreboard

The headline numbers Artemis-Team agents quote everywhere — presentations, social, billboards, listing appointments. Filtered by your selected view above.

Agent Baseball Cards

The people behind the numbers

Each card shows that agent's signature accomplishments. Click More Stats to see their full library — veterans have 10–20 signature stats, rising agents grow into more as their book builds.

Show stats from:
Methodology

Where the numbers come from

Every stat traces back to one of three foundations.

A · Live MLS Data

93,728+ closed residential transactions in MARIS since 2024. 8+ years of historical depth (2018-onward). Direct PostgreSQL replica with field-level access to listing details, marketing inputs, and transaction outcomes. Bi-state coverage (IL + MO).

B · Peer-Reviewed Research

49 academic papers anchor the methodology: Knight 2002 (Double Penalty of overpricing), Han-Strange 2014 (Bidding-war premium), Levitt-Syverson 2008 (informed-asymmetry baseline), plus 46 more spanning negotiation, pricing, and consumer outcomes.

C · Internal Artemis Records

381 transactions annualized in 2026 (target 500). $100M+ team volume rolling-12. Per-agent baseball cards built from individual transaction histories validated against MARIS records since 2018.

How each metric is computed — full per-stat methodology (v1.2)

Scorecard version: v1.2 (per-cohort regression, 7 metric families × 4 time periods). Refreshes when MARIS data refreshes. Every per-agent cell carries a publishable flag — if the sample size is below threshold, the stat is silently hidden from the public side.

1 · Seller Premium Added (Agent WAR)

Formula: (Close Price − Predicted Value) − DOM-drag − Concessions, where Predicted Value = per-cohort linear regression α + β × sqft fit per (city × year × price-bracket) cohort with n ≥ 5 (Option A). Falls back to median $/sqft when regression is degenerate.

Captures "wins above replacement" — dollars the agent added vs. what a cohort-typical agent would have produced on the same home. Published when n ≥ 15.

2 · Buyer Navigation Impact

Formula: (Original List Price − Close Price) + Concessions captured, averaged across buyer-side closings.

Captures total buyer-side dollar value extracted — both negotiated price reductions and concession capture. Published when n ≥ 10.

3 · Buyer Bidding-War Win Rate over-ask proxy

Formula: (buyer-side closings where Close Price ≥ Original List Price) / (all buyer-side closings) × 100.

When a buyer-side closing went at or above the seller's original ask, that's ~95% likely a bidding war the buyer won — sellers don't take above-ask offers outside competitive situations. We can't see losing offers from MLS data alone, so this is a proxy, not a direct measurement. Published when n ≥ 10.

Tightening this signal — counting every offer written (winners + losers) for a true win-rate denominator — requires DotLoop ingestion. Separate workstream, in progress.

4 · Over-Asking Sale Rate (listing side)

Formula: (this agent's listings closed > Original List Price) / (all this agent's closings) × 100.

The listing-side mirror of #3 — how often this agent's marketing engine produces a bidding war the seller benefits from. Published when n ≥ 10.

5 · Avg Below-List Savings

Formula: AVG(Original List Price − Close Price) on buyer-side closings where Close < List.

Pure negotiation extraction — how many dollars the agent pulled OFF the price when the market allowed. Published when n ≥ 10 below-list buys.

6 · Median DOM on Winning Offers (buyer side)

Formula: Median of Cumulative Days on Market (CDOM) across buyer-side closings.

Speed signature — agents who consistently close fast on shorter-DOM listings are reading the market in real time. Published when n ≥ 10.

7 · Concession Capture Rate

Formula: (buyer-side closings with concessions > 0) / (all buyer-side closings) × 100, plus avg $ when captured.

How frequently the agent extracts concessions on behalf of buyers — separate from price negotiation, this is closing-cost help, repair credits, points, etc. Published when n ≥ 10.

8 · Listing Completion Rate

Formula: closed / (closed + expired + canceled + withdrawn) × 100.

Industry-wide baseline is ~65–75%. Higher = agent finishes what they list, not just collects listings. Published when n ≥ 10 terminal listings.


Time periods

Every metric is computed across four parallel cuts: Lifetime (career to date), Since 2024 (the Artemis modern-era cohort), 2025, and 2026 YTD. Year-over-year inflections are visible at a glance.

Exclusions

Bi-state IL+MO residential only. Listings < $50K original list price excluded (data errors). Close-to-list ratios outside 0.5–2.0× excluded (commercial flips, distressed). Tear-downs excluded when close $/sqft is < 65% of cohort median (different asset class).

Why per-cohort regression (Option A)

The earlier v0.6 methodology used a flat sqft × median-$/sqft baseline. It systematically under-predicted large homes (the "Kelly Paradox") because price doesn't scale linearly with size. v1.1 fits a fresh regression per (city × year × bracket) cohort to capture the non-linear shape. v1.2 preserves that core and adds time-period granularity + four new buyer-side metric families.

Want to be on this scoreboard?

The Artemis Team is hiring producers who want measurable, defensible numbers behind their name. If you want to know what it'd take to make this list, tell us about yourself.