Q2 2026 accuracy tracker: Krevmire-area licensed prophets
Quarterly tracking post. Methodology same as always: only counting prophecies with verifiable outcomes that resolved within the quarter. Partial fulfillment counted as 0.5. Self-contradicting / bifurcated prophecies excluded from the dataset. Sample sizes in parentheses. Madame Thessaly (Crale St) ......... Q2: 68% (n=22) Q1: 54% (n=19) up Orren Voss (Municipal Desk) ........ Q2: 73% (n=41) Q1: 71% (n=38) up The Fenwick Seer ................... Q2: 81% (n=11) Q1: 84% (n=9) down Glimma Datch (by appt) ............ Q2: 59% (n=8) Q1: 62% (n=12) down PropheCo Automated Kiosk (GrelMart) Q2: 44% (n=67) Q1: 46% (n=71) down Notes: Thessaly's spike is interesting. She may have changed methodology or she may just be having a good quarter. The Fenwick Seer remains the accuracy leader but his sample size is tiny because he only takes clients "when the corridor allows it," which in practice means you show up and hope he's in the mood. PropheCo kiosk continues to perform below coin-flip accuracy, which is statistically impressive in its own way.
Thessaly's Q1 numbers were dragged down by that batch of career prophecies that all resolved late. If you adjust for timing lag she's been in the high 60s consistently. Also, she doesn't charge for re-reads if the prophecy doesn't resolve within the stated window, which PropheCo definitely does not offer. The kiosk charged me twice for a prophecy about my parking situation that turned out to be about someone ELSE'S parking situation.
Are you counting the Fenwick Seer's weather prophecies or just life readings? Because his weather accuracy is near 100% but it's widely suspected he's just looking outside before clients arrive. If you strip weather from his dataset his number drops a lot.
Fair point. I've been including weather because it's technically a verifiable outcome but I'll run the numbers without it for Q3. If he's gaming the dataset by frontloading weather readings I want to catch that.
In defense of PropheCo: the kiosk at GrelMart is the only prophetic service available 24/7, costs 8 credits per reading, and doesn't require an appointment. At 44% accuracy you're getting slightly-less-than-random guidance at any hour of the night for the cost of a sandwich. I used it at 2 AM last month to decide whether to text my ex. It said "the winds of change blow toward familiar harbors." I texted her. She did not respond. I'm counting that as a correct prediction because the wind blew and nothing changed, which is familiar.
I appreciate the defense but I cannot in good conscience count "my ex didn't text back and I decided that counts" as a verified outcome. The kiosk's real problem isn't accuracy per reading, it's that the volume is so high it drags the average into statistical visibility. When you deliver 67 prophecies in a quarter, 44% means 37 misses. Thessaly's 68% on 22 readings is 7 misses. People notice 37 misses. They talk about 37 misses. At the Crambleside weekend market there is a woman selling PropheCo complaint forms as a novelty item and they are outselling the actual novelty items.