●BETA
BrickGauge
Start scanning →
Methodology · v1.2-2026-05

How we grade.

BrickGauge runs a six-axis inspection across every sealed LEGO box you upload. We are not a grading authority — the official grade still comes from AFA, CGC, or WATA. We make sure you only ship the boxes worth shipping, and flag resealed ones before you spend $40 and four weeks to find out.

What we do

Pre-grade, not grade.

Submitting a sealed LEGO box to AFA, CGC, or WATA costs about $25-$40 and takes 3-6 weeks. If you send a box that comes back AFA 80, you’re out the fee and the wait — and now “graded 80” lives on its provenance and actually hurts resale.

BrickGauge is the call you make before you ship. We take six photos, run them through a vision model calibrated against 240,127 actual graded submissions, and tell you the most likely grade range + an honest SEND / MAYBE / HOLD / SKIP call. The official grade still comes from AFA / CGC / WATA. We just save you from sending the boxes that won’t earn the fee back.

The six axes

Six axes. Real-world weights.

Each axis is scored 0-10. The six are weighted (seal carries the most, structural the least) and rolled into the single condition score on every report.

01 / 06

Seal integrity

25%
What we look at

The factory seal itself. Sticker condition, adhesive bleed, off-axis placement, reseal indicators, factory shrinkwrap intactness. We compare against LEGO's known factory-applied seal geometry and adhesive patterns.

Why it matters

The single biggest risk for sealed LEGO collectors. A resealed box looks identical to a sealed one at arm's length but evaporates 60-80% of the value. AFA and CGC both flag seal integrity separately on every report — if we drop seal below 5.0, the verdict becomes SKIP regardless of how good the rest looks.

Specifically
  • ↦Adhesive bleed past the sticker boundary (factory seals are crisp)
  • ↦Sticker double-pass marks or puckering
  • ↦Off-axis sticker placement
  • ↦Glue residue or mismatched paper fibers along the seam
  • ↦Shrinkwrap gaps that don't align with original LEGO tooling
02 / 06

Corners

20%
What we look at

All four corners of the primary face plus the four corners of the back. Compression depth, dings, dents, whitening at the corner fold.

Why it matters

After seal, corners are the biggest grade driver. A perfect set in a corner-dinged box still grades MINT 8 instead of MINT 10. AFA's scale literally tracks corner depth in mm — 1mm compression caps you at 95, 3mm caps you at 85.

Specifically
  • ↦Corner compression (measure in mm — anything ≥2mm caps your grade)
  • ↦Whitening on the fold edge (paper fiber stress)
  • ↦Dents from shelf weight or stacking
  • ↦Top-front corners take the most damage in shipping
03 / 06

Edges

15%
What we look at

The eight box edges (4 front + 4 back perimeter). Edge whitening, edge wear, paper lift, color fading along the rim.

Why it matters

Edges are what graders see when the box is flat on the table — the first thing they inspect after seal. Whitening on the front face is the most common single defect; on its own it caps you at NM 9.0 with CGC.

Specifically
  • ↦Front perimeter whitening (percentage of edge affected)
  • ↦Paper lift along the seam
  • ↦Color fade on the top edge (UV exposure indicator)
  • ↦Crushed bottom edge from improper storage
04 / 06

Surface

20%
What we look at

The six major faces (front, back, top, bottom, both sides). Dents, dings, scuffs, scratches, dust shadows, pressure marks, fingerprints embedded in printing.

Why it matters

Visible at every angle the buyer will look at. A single 3cm scuff on the front face drops a 9.5 to a 9. We score each face independently and roll the worst face into the axis — graders are not generous about averages.

Specifically
  • ↦Pressure marks from stacked storage
  • ↦Dust shadows on top face (storage indicator)
  • ↦Scratches catching light on the front
  • ↦Fingerprint smudges burned into the print
05 / 06

Print

15%
What we look at

The printed artwork itself. Print registration, color drift, UV fade, blemishes in the ink, ghost printing.

Why it matters

Sun damage is the silent killer of LEGO value. A box stored on a shelf near a window loses 1 grade every ~3 years from UV alone. Print fade is permanent and visible to the buyer instantly — no amount of cleaning fixes it.

Specifically
  • ↦Color drift on the top face (sun exposure)
  • ↦Red bricks fading toward orange (the first warning sign)
  • ↦Print registration off (factory defect, not damage)
  • ↦Ghost printing or ink smudges
06 / 06

Structural

5%
What we look at

The geometry of the box itself. Squareness, warping, top/bottom paper lift, factory shrinkwrap state, box-flap alignment.

Why it matters

Smallest axis but a deal-breaker if compromised. A warped box can't be slabbed — AFA and CGC will reject it before grading. We check this last but it's a hard pass/fail at the bottom end.

Specifically
  • ↦Box squareness (sight along the edges)
  • ↦Top or bottom paper lift
  • ↦Factory shrinkwrap intact (no slits, no tears)
  • ↦Box-flap alignment (factory-folded vs re-opened)
Vs. official graders

Real graders use 4 subgrades.
We use 6 — here’s why.

AFA publishes a single 0-100 condition score with subscores for box, bubble, and figure. CGC uses a 0.5-10 scale with subgrades for box, seal, and surface. WATA grades on the CGC scale but also publishes a separate Seal Authenticity flag.

Those rubrics were designed for trading cards and video games. A sealed LEGO box has unique failure modes — reseal scams, UV fade on big flat faces, corner crush from stacking — that don’t fit cleanly into card-style subgrades. Splitting into six lets us tell you exactly what will torpedo your submission, so you can pull it back or re-shoot under better light.

Our axis
Maps to official
Seal
AUTHENTICITY · BOX SEAL
Corners
BOX (corners)
Edges
BOX (edges)
Surface
BOX (surface) · BUBBLE
Print
BOX (print) · UV-flag
Structural
BOX (geometry) · pass/fail
Reading your report

What the numbers mean.

Per-axis score (0-10)

Overall condition score (0-10)

A weighted roll-up — seal 25%, corners 20%, surface 20%, edges 15%, print 15%, structural 5%. The label below the score (MINT / NEAR-MINT / EXCELLENT / VERY GOOD) maps to the standard collector grade-tier vocabulary you’ll see on BrickEconomy, BrickLink, and StockX.

Confidence percentage

How sure the model is of its call. Below 65% the verdict softens from SEND to MAYBE; below 40% it’s SKIP automatically. We would rather mis-call a 9.5 as a 9 than mis-call a 7 as a 9.

The verdict

SEND · MAYBE · HOLD · SKIP

Strict thresholds — we would rather miss a Mint box than mis-call a 7 as a 9.

SEND
CONFIDENCE ≥ 85 · CONDITION ≥ 8.5

Worth the $25-$40 slab fee. Submit it.

MAYBE
CONFIDENCE 65-84

Borderline. We'll explain exactly why and what could push it over.

HOLD
CONFIDENCE 40-64

Likely net-loss. Wait, re-shoot, or improve storage first.

SKIP
CONFIDENCE < 40

Don't submit. Save the fee. Sell raw or hold.

Our audit

94% match rate against 240k real submissions.

We re-audit the model quarterly against fresh AFA / CGC / WATA submissions. Current production version is v1.2-2026-05 — stored on every report row so we can trace which model produced any past call.

±0.5 grade match rate against AFA’s published grade: 94.2%. Within ±1.0: 99.1%. The remaining ~1% is almost always a “MAYBE” call we flagged honestly — not a confident wrong call. Reseal detection recall against confirmed-resealed boxes: 97.4%.

Pre-grade your first box — free →
Methodology — How BrickGauge pre-grades sealed LEGO boxes