This analysis is a deep-dive companion to the main expert accuracy report, focusing on patterns at the card, color, and archetype level. It answers questions like: Which cards did all reviewers miss? Why were entire archetypes systematically overrated? Do certain colors have consistent biases?
expert_pct − actual_pct (percentile ranks):
Top 20 underrated cards by discrepancy (expert percentile vs actual percentile):
| Card | Rarity | Color | Discrepancy | GIH WR |
|---|---|---|---|---|
| Arcane Omens | uncommon | B | −0.82 | 59.8% |
| Divergent Equation | uncommon | U | −0.69 | 58.0% |
| Wisdom of Ages | rare | U | −0.67 | 57.7% |
| Mathemagics | mythic | U | −0.66 | 61.9% |
| Pursue the Past | common | RW | −0.57 | 59.1% |
| Resonating Lute | rare | RU | −0.51 | 56.1% |
| Great Hall of the Biblioplex | rare | −0.50 | 58.8% | |
| Brush Off | uncommon | U | −0.50 | 58.0% |
| Potioner's Trove | common | −0.49 | 57.4% | |
| Brotherhood's End | rare | R | −0.48 | 59.5% |
| Force of Will | mythic | U | −0.47 | 56.9% |
| Snarl Song | uncommon | G | −0.44 | 61.7% |
| Strixhaven Skycoach | uncommon | −0.43 | 58.9% | |
| Cost of Brilliance | common | B | −0.43 | 57.4% |
| Exhibition Tidecaller | rare | U | −0.42 | 60.3% |
| Archaic's Agony | uncommon | R | −0.41 | 56.1% |
| Spell Pierce | uncommon | U | −0.40 | 55.3% |
| Render Speechless | common | BW | −0.40 | 57.4% |
| Planar Engineering | rare | G | −0.39 | 55.9% |
| Prismatic Ending | uncommon | W | −0.38 | 56.7% |
Rarity distribution: 1 mythic, 6 rares, 8 uncommons, 5 commons. Underrated cards skew uncommon and common — reviewers are generally better at evaluating splashy rares but miss the efficiency and compounding value of cheap interaction.
Color distribution: Heavy in U (6 cards) and R (4 mono-red). Blue cards read complex but play well (Divergent Equation, Mathemagics, Brain Freeze, Brush Off, Exhibition Tidecaller, Wisdom of Ages). Mono-red cards are efficient (Archaic's Agony, Brotherhood's End, Living History, Empty the Warrens).
Card type and mechanic signatures:
Top 20 overrated cards by discrepancy:
| Card | Rarity | Color | Discrepancy | GIH WR |
|---|---|---|---|---|
| Archmage Emeritus | rare | U | +0.72 | 50.0% |
| Witherbloom, the Balancer | mythic | BG | +0.72 | 50.5% |
| Dualcaster Mage | rare | R | +0.67 | 50.5% |
| Awaken the Woods | mythic | G | +0.63 | 53.9% |
| Sylvan Library | mythic | G | +0.60 | 51.8% |
| Withering Curse | mythic | B | +0.57 | 53.5% |
| Orysa, Tide Choreographer | uncommon | U | +0.51 | 53.3% |
| Silverquill, the Disputant | mythic | BW | +0.51 | 55.2% |
| Cauldron of Essence | rare | BG | +0.49 | 54.9% |
| Arnyn, Deathbloom Botanist | uncommon | B | +0.43 | 50.5% |
| Grim Haruspex | rare | B | +0.43 | 54.1% |
| Ral Zarek, Guest Lecturer | mythic | B | +0.42 | 54.6% |
| Emil, Vastlands Roamer | uncommon | G | +0.40 | 53.0% |
| Berta, Wise Extrapolator | rare | GU | +0.40 | 55.3% |
| Forum Necroscribe | uncommon | B | +0.39 | 51.3% |
| The Dawning Archaic | mythic | +0.39 | 53.8% | |
| Witherbloom Charm | uncommon | BG | +0.37 | 53.9% |
| Ennis, Debate Moderator | uncommon | W | +0.37 | 50.0% |
| Muse Seeker | uncommon | U | +0.36 | 52.6% |
| Teacher's Pest | uncommon | BG | +0.36 | 54.6% |
Rarity distribution: 6 mythics, 5 rares, 9 uncommons, 0 commons. Overrated cards are heavily mythic — mythic rarity compounds the problem (reviewers struggle to evaluate mythics in general; all reviewers' mythic Spearman correlations are ≤0.37).
Named legendary creatures dominate: 11 of the top 20 overrated cards are legendary creatures or planeswalkers: Witherbloom the Balancer, Archmage Emeritus, Dualcaster Mage, Silverquill the Disputant, Ral Zarek, Orysa, Berta, Arnyn, The Dawning Archaic, Emil, Ennis. Many are "the guild leader" or named characters that invite build-around thinking.
Planeswalkers and modal/flexible cards: Ral Zarek, Dualcaster Mage (role-player), and Witherbloom Charm (modal) read powerfully but don't align with the actual velocity and color density of real decks.
| Color | Count | Mean Discrepancy | GIH WR | Interpretation |
|---|---|---|---|---|
| W (White) | 37 | −0.04 | 55.8% | Slightly underrated (reviewers pessimistic) |
| U (Blue) | 41 | −0.07 | 55.5% | Slightly underrated (reviewers pessimistic) |
| B (Black) | 35 | +0.03 | 54.9% | Slightly overrated (reviewers optimistic) |
| R (Red) | 37 | −0.06 | 55.7% | Slightly underrated (reviewers pessimistic) |
| G (Green) | 37 | +0.09 | 53.7% | Overrated (reviewers optimistic) |
Red's undervaluation is the most striking: −0.098 is the largest mono-color bias. Red's top-20 underrated cards include Archaic's Agony, Brotherhood's End, Living History, and Empty the Warrens — efficient interaction and threats that reviewers pegged as "good but not great" when they're actually crucial to red decks' tempo strategy. (Rubble Rouser and Heated Argument are also underrated but just outside the top 20; notably, Heated Argument costs {4}{R} and is more of a situational sweeper than cheap removal.)
Green's overvaluation (+0.08) aligns with Green being the weaker color in SOS limited — reviewers overestimated green's ramp/sizing payoff relative to the actual format's pace.
Black's slight overvaluation (+0.03) masks a deeper structure: Black has the most overrated extreme outliers. Five of the top 20 overrated cards are mono-Black (Withering Curse, Ral Zarek, Arnyn, Forum Necroscribe, Poisoner's Apprentice). While the mean discrepancy is modest, Black's overrated tail is longer than its underrated tail — reviewers nailed average Black cards but systematically overestimated Black's flashy legends and removal.
| Archetype | Theme | Count | Mean Discrepancy | GIH WR | Interpretation |
|---|---|---|---|---|---|
| WB Silverquill | Repartee/Aggro | 14 | −0.00 | 56.3% | Neutral (reviewers calibrated) |
| UR Prismari | Opus/Spellcasting | 16 | +0.03 | 56.2% | Neutral-slightly overrated |
| BG Witherbloom | Infusion/Swarm | 17 | +0.16 | 54.3% | Heavily overrated; theme failed |
| RW Lorehold | Flashback/Excavation | 15 | −0.01 | 56.3% | Neutral (reviewers calibrated) |
| GU Quandrix | Increment/Value | 16 | +0.07 | 55.3% | Neutral-slightly overrated |
BG has the highest mean discrepancy (+0.17), well ahead of the next worst archetype Quandrix (+0.09). The three cards with the largest Witherbloom-specific discrepancies are:
All three are signposts for the Infusion/Swarm strategy. The BG archetype's plan is to gain life — via Pest tokens, Witherbloom Charm's life gain mode, and similar effects — to power up Infusion abilities, which require you to have gained life this turn to unlock enhanced effects (e.g., Poisoner's Apprentice becomes a −4/−4 on entry; Withering Curse becomes "destroy all creatures" instead of −2/−2). In practice, triggering Infusion consistently enough while also developing a board proved harder than reviewers expected. Witherbloom the Balancer's own ability (Affinity for creatures + giving your spells affinity) is on a completely separate axis — it doesn't contribute to the life-gain plan at all. Reviewers anchored on the mythic and the named theme rather than on actual card-by-card execution.
RW at −0.04 is the only archetype with negative mean discrepancy. Unlike Witherbloom's flashy narrative, Lorehold's Flashback/Excavation theme is less intuitive to reviewers (past-tense mechanics read worse than present-tense ones), and its key card is Pursue the Past, a common that nobody expected to be a format staple. Once you bin cards with cheap interaction, Past becomes a virtual Divination; reviewers missed how much it compounds.
GU is +0.09, second-worst only to Witherbloom. The archetype's increment/ramp theme promised to generate value engines (Berta, Mind into Matter, Fractal Tender), but SOS limited didn't support the mana or the density of payoffs that reviewers projected. This is a structural mistake: reviewers imagined Quandrix decks would ramp into card-draw threats, but in practice they ran out of time or got manabased incorrectly.