We Roasted Thousands of Outfits: The 5 Mistakes That Got the Lowest Scores
When AI stops being polite and starts being brutally honest, patterns emerge fast. The lowest-scoring roasted outfits — some scoring as low as 5 out of 100 — all share the same five mistakes. Here's what the AI said, word for word, and what to do instead.
Published: April 21, 2026
Reading time: 7 minutes
Roast mode doesn't round up. The average roasted outfit scores 30 out of 100. The lowest recorded score is 5 — and the AI explained exactly why.
Standard outfit analysis is calibrated to be constructive. It finds the strengths, acknowledges the effort, and delivers improvements diplomatically. Roast mode is different. It evaluates the same five dimensions — colour harmony, fit, details, cohesion, and occasion-appropriateness — but applies no diplomatic softening to the feedback or the score.
The result is a much lower average score (30 vs 60 for standard analysis) and feedback that is direct, specific, and occasionally uncomfortable. After analysing thousands of outfit photos, including roast-mode submissions, we identified the five mistakes that consistently produce the lowest scores. They're listed here in order — from most common to most severe.
Before you read: a note on roast mode
Roast mode is optional and consent-based. You choose the severity — Gentle Shade, Full Roast, or No Mercy. The AI feedback quoted in this article is from Full Roast and No Mercy submissions. If you want a kinder verdict, standard outfit analysis is available anytime.
Mistake #1: Wearing Loungewear in Public
The most frequently flagged mistake across all roast-mode analyses. Not "casual" outfits — actual loungewear: pyjama sets, fuzzy robes, sleep shorts with hoodies, or tracksuit sets worn with house slippers. The AI does not treat this as a matter of personal style. It calls it what it is.
ROAST VERDICT (No Mercy)
"This is currently pajamas/loungewear. Do not wear this outside. This look is strictly limited to domestic environments; it lacks the necessary polish for social outings."
The score impact is significant. Occasion-appropriateness — which averages 14.4/20 in standard analysis — drops to near-zero for loungewear worn in a public context. No amount of styling fixes that.
The fix: If you love the comfort of loungewear, transition the pieces individually. Joggers with a structured tee and clean sneakers is not loungewear. A hoodie with fitted jeans and a leather jacket is not loungewear. The problem is the combination, not the individual items.
It's not the hoodie. It's not the joggers. It's wearing them together with slides that signals no sartorial awareness whatsoever.
Mistake #2: Graphic Print as the Entire Outfit
A single graphic piece — a statement tee, a bold print hoodie — can work. Making it the anchor of the entire outfit with nothing to counterbalance it is what the AI flags. The problem compounds when the graphic print is also dated (early-2000s brand logos, faded text, outdated character graphics).
ROAST VERDICT (Full Roast)
"The graphic sweatshirt significantly lowers the maturity of the overall aesthetic. The graphic print on the sweatshirt feels dated and the denim wash lacks character or premium texture. The outfit lacks accessories or layering depth, appearing as a collection of random basics rather than a curated look."
The fix: One of three approaches. Replace the graphic piece with a solid-colour version of the same garment. Keep the graphic piece but surround it with intentional solids and add structure through outerwear. Or wear the graphic piece tucked in with a belt so it reads as an accent rather than the entire style statement.
Mistake #3: Indoor Footwear Worn Outside
This one is specific and non-negotiable in AI analysis. Fuzzy slides, house slippers, flip-flops worn with non-beach outfits, and socks without shoes all trigger explicit style warnings. The AI doesn't soften this.
ROAST VERDICT (No Mercy)
"Fuzzy slides are strictly for indoor use; wearing them outdoors will immediately degrade the perception of you and the entire outfit. Wearing socks without footwear in a public setting is a major styling error."
The fix: Clean white leather sneakers or tan leather slides (not fuzzy — leather) are the universal casual footwear fix. Chelsea boots for any outfit that needs elevation. The footwear sub-category is the highest-impact single accessory in outfit analysis — getting it right is worth 7 points on average.
All five mistakes are entirely avoidable. The lowest-scoring outfits in the data share at least three of these simultaneously.
Mistake #4: Zero Waistline Definition on a Shapeless Silhouette
An untucked oversized top over loose-fitting bottoms is the silhouette that roast mode reserves its harshest fit-dimension scores for. Not because oversized is wrong — but because shapeless without intent reads as unaware, and AI roast mode is specifically calibrated to call out unawareness.
ROAST VERDICT (Full Roast)
"The current silhouette is extremely boxy; consider a half-tuck or a more fitted bottom to balance the volume. The outfit is appropriate for casual lounging, though it lacks the polish required for public social settings."
The fix: French tuck the front of the top. Add a leather belt at the waist. Or switch one volume piece — if the top is oversized, go fitted on the bottom, or vice versa. The rule is: one piece can be relaxed, not both simultaneously.
Mistake #5: Absolutely No Accessories
A standard outfit analysis flags missing accessories as the most common improvement area. Roast mode doesn't flag it — it points at it directly. An outfit with zero accessories — no jewellery, no belt, no structured bag, no watch — scores near-zero on the styling and cohesion dimension in roast mode, because there is no evidence of a styling decision beyond putting on clothes.
ROAST VERDICT (No Mercy)
"The outfit lacks accessories or layering depth, appearing as a collection of random basics rather than a curated look. This look is strictly limited to casual errands or low-key social outings. Zero visual interest has been created."
The fix: Add one. Just one. A silver chain necklace, a minimalist watch, a leather belt, or a pair of gold hoop earrings. The difference between a 28-scoring roasted outfit and a 52-scoring one is sometimes literally a single £12 chain necklace. The AI doesn't require perfection — it requires evidence of intent.
The Common Thread: Intentionality
Every one of these five mistakes signals the same thing to AI analysis: the absence of a deliberate styling decision. Loungewear outside signals no consideration of context. A full graphic-print outfit signals no thought about visual balance. Indoor shoes outside signal no awareness of footwear's role. Zero waistline definition signals no silhouette management. No accessories signal no finishing of the look.
AI roast mode is specifically calibrated to be unsparing about this. Standard outfit analysis gives partial credit for intent — "appropriate for casual settings" — and that's how a 45-point outfit becomes a 60. Roast mode doesn't extend that credit. If the intent isn't visible in the photo, it doesn't exist in the score.
The good news: intentionality is the cheapest styling improvement possible. A French tuck is free. A silver chain is £12. Wearing outdoor shoes outdoors is also free. The five mistakes above are all rooted in decisions, not resources.
Think your outfit can survive a roast?
Choose your severity level — Gentle Shade, Full Roast, or No Mercy — and let the AI tell you exactly what it really thinks. No diplomatic softening.
Roast My Outfit 🔥 Standard Analysis →Frequently Asked Questions
What is an AI outfit roast?
An AI outfit roast is a brutally honest fashion analysis with no diplomatic softening. OutfitScore's Roast My Outfit feature offers three severity levels: Gentle Shade, Full Roast, and No Mercy. The same five scoring dimensions apply, but the feedback is direct and the score is not rounded up for effort.
What is the lowest possible outfit roast score?
The lowest recorded score in our data is 5 out of 100. The average roast score is approximately 30 — roughly half the standard outfit analysis average of 60. Roast mode does not apply diplomatic adjustments to scores.
Is roast mode harsher than standard outfit analysis?
Yes, significantly. Standard analysis is calibrated to be constructive and balanced, averaging around 60/100. Roast mode averages around 30/100 and uses direct language without softening. The criteria are identical, but the interpretation and feedback tone are completely different.
Can a roasted outfit still have good advice in it?
Yes — the roast always includes improvement tips alongside the harsh feedback. The severity of the criticism doesn't affect the quality of the advice. Often, the most useful outfit feedback comes from roast mode precisely because it's specific and unsparing.