We Tested Our Reprint Risk Scores Against Real Commander Reprints. Here's What We Found.

We talk a lot about reprint risk on SpellBook Finance. It's one of the biggest factors in MTG investing — a single reprint announcement can shave 30-60% off a card's value overnight. So when we built our reprint risk algorithm, we knew we'd eventually have to test it against reality. That's what this post is about.

The Test

We took our reprint risk scores and checked them against 45+ cards that were actually reprinted in the Strixhaven Commander preconstructed decks. These precons are a perfect test case: Wizards of the Coast selects cards across a wide price and popularity range, and the decklists are public record. No cherry-picking — every card in the precon gets checked against what our algorithm predicted.

The result: roughly 65-70% useful signal. Our algorithm correctly flagged the majority of reprinted cards as elevated risk before the reprint happened. Not bad for a first pass, but the 30-35% it missed told us more than the ones it got right. Those misses clustered into three specific blind spots, and finding them is what led to the v2 model.

You can explore current reprint risk scores on the card screener.

Three Blind Spots We Found

1. The price dead zone ($5-$20)

Our v1 algorithm treated price as a linear signal: higher price meant higher reprint risk. Sounds reasonable, but it created a blind spot right in the sweet spot that WotC loves for Commander precons. Cards in the $5-$20 range were being scored as relatively safe because they weren't expensive enough to trigger the high-price risk factor. In reality, this is exactly the price range WotC targets most aggressively — expensive enough to be a selling point on the back of the box, but cheap enough that reprinting them doesn't tank an entire set's value.

The fix: we replaced the linear price signal with a bell curve that peaks in the $5-$20 range. Cards in this sweet spot now correctly receive elevated likelihood scores. Cards under $2 (not worth the slot) and cards over $80 (WotC is more cautious with premium reprints) score lower on the price-driven likelihood component.

2. No reprint velocity signal

Cards like Ghostly Prison and Three Visits have been reprinted many times. Our v1 algorithm actually treated high printing counts as a safety signal — the logic being that a card with 15 printings had already been reprinted enough. Wrong. Cards that get reprinted frequently get reprinted frequently because WotC considers them staple inclusions. A card reprinted twice in the last three years is more likely to be reprinted again than a card with the same total printing count that hasn't seen a new version in five years.

The fix: we added a trailing 36-month reprint velocity metric. Cards reprinted 2+ times in the past three years now receive a +15 boost to their likelihood score. This correctly captures cards that WotC has on their "go-to reprint" list.

3. A single blended score

This was the deepest problem. Our v1 model produced one number — a composite reprint risk score from 0 to 100. But that single number was conflating two fundamentally different questions: "Will this card be reprinted?" and "How much will I lose if it is?" A $3 Commander staple and a $50 Commander staple could receive the same composite score because one had high likelihood but low financial impact, while the other had moderate likelihood but devastating financial exposure. The single number made them look equivalent when the correct investment response to each is completely different.

The Fix: A Dual-Score Model

Version 2 splits reprint risk into two independent axes:

Likelihood — the probability that WotC will reprint this card. Factors include:

• Total printing count (more printings = WotC is comfortable reprinting it)
• Years since last printing (longer gap = lower near-term likelihood)
• EDHREC popularity (high-demand Commander staples are reprint magnets)
• Reprint velocity (2+ reprints in 36 months = WotC has this card on rotation)
• Universes Beyond status (UB cards have different reprint economics)

Impact — how much value you stand to lose if the reprint happens. Factors include:

• Current price (full weight — a $50 card losing 40% hurts more than a $5 card losing 40%)
• EDHREC popularity at 25% weight (high-demand cards have a larger "blast radius" of affected portfolios)
• Rarity (mythics tend to hold value better post-reprint than rares or uncommons)

The two scores combine into a 2x2 category matrix that tells you not just the risk level, but what to do about it:

• Ticking Clock (high likelihood + high impact) — sell or trade before the reprint lands. These are the cards most likely to cost you real money.
• Glass Cannon (moderate likelihood + high impact) — the reprint may not be imminent, but if it happens, the loss is significant. Monitor closely and have an exit plan.
• Phoenix (high likelihood + low impact) — will probably be reprinted, but the financial hit is manageable. These cards tend to recover quickly because demand stays high.
• Fortress (low likelihood + low impact) — Reserved List cards and similar. Minimal risk on both axes.
• Wait and See (moderate on both axes) — not urgent, but worth checking back after the next reprint wave.

You can filter by these categories on the screener page.

Before and After: Three Examples

The $8 Commander staple

Under v1, a card like this scored around 45 — a Grade C, "moderate risk." The algorithm saw the sub-$10 price and treated it as relatively safe. Under v2, it's correctly categorized as a Ticking Clock: high likelihood (it's a popular Commander card in WotC's favorite reprint price range) with moderate impact (you won't lose a fortune, but across a playset or a stack of copies, it adds up). The $5-$20 sweet spot is where WotC reprints most aggressively, and the algorithm now reflects that.

The $50+ mythic with one printing

Under v1, this scored near 100 — a Grade F, maximum risk. The high price drove the score through the roof. Under v2, it's a Glass Cannon: moderate likelihood (mythic rarity is harder for WotC to slot into precons and supplemental products) but high impact (if it does get reprinted, you're losing significant value). The category name tells you what to do — this card is fragile, and you should have a plan for when it breaks.

The Reserved List card

Under both v1 and v2, this is a Fortress. Zero reprint likelihood, and the algorithm correctly identifies it as such. The Reserved List is the one case where our v1 model was already right, and v2 didn't need to change anything.

What's Next

Transparency is a feature. We're not going to publish an algorithm and then never tell you how it performs. We're actively tracking reprint velocity data across every Commander release, supplemental product, and Secret Lair drop to continuously calibrate the model.

After the next major Commander precon release, we'll run this same test again and publish the results. If accuracy improves, great — we'll show the numbers. If we find new blind spots, we'll document those too and explain how we're fixing them. That's how you build trust with data.

If you want personalized reprint risk analysis for your specific portfolio, the SpellBook Copilot can walk you through your highest-exposure holdings and suggest actions based on the dual-score model.

Reprint risk scores are algorithmic estimates based on historical data and card metadata. They do not constitute financial advice. Wizards of the Coast reprint decisions are ultimately unpredictable — our goal is to give you the best available signal, not certainty.