5/3/2025, 3:52 PM: ELO Sorting My Baseball PC
After a full morning elbow-deep in rankings, matchups, and tuning logic, I’ve landed on a system that’s finally doing what I need: making sense of my personal card collection using a dynamic, living rating model.
Why ELO?
I’m not trying to create some universal “value” system. This isn’t about market comps or PSA 10s. This is about how I feel about my cards—relative to each other. Sentimental weight, aesthetics, print run quirks, storylines—all of it can be surfaced if the system allows for fluid context. That’s what ELO does better than tier lists or 1–10 scoring. It gives each card the chance to rise, fall, surprise, or age like wine—based on real head-to-head outcomes.
The Current State
We now have 32 cards, each with its own rating, match count, and an assigned tier (S through D). Elo is calculated via a custom Google Sheets script, using a standard kFactor = 32, and ratings update automatically when matchups are marked "done".
The best cards right now:
Aaron Judge T210, Lindor Blue Atomic, and Nimmo Snapshots Auto — all sitting solidly in S-Tier with 10–15 matchups under their belts.
At the bottom:
Tony Gwynn Chrome, Vladdy Jr. Bowman, and J.P. Morgan Heritage have sunk to the D-Tier after repeated trials.
And in the middle? That’s where the most interesting movement happens. A-Tier and B-Tier are volatile, still tuning themselves.
Process Design
After some experimentation, I’ve adopted a two-phase pipeline for evaluating cards:
1. Stagger Phase (New Card Intake):
When I add new cards:
They all start at 1000.
Each gets 3 initial matchups: the lowest A/S-tier, the first B-tier above 1000, and the highest D-tier.
This gives me a fast read without exhausting the whole list.
2. Tuning Phase (Pressure Testing):
Once those cards have ~3 matchups, I move them into tuning:
1 matchup vs a higher card in the same tier
1 matchup vs someone ±1 tier
1 matchup vs their tier floor
This gets them to 5–6 total matchups, which is the sweet spot for Elo to settle. At that point, I trust where they are and can shift focus to others.
Why This Matters (To Me)
This isn’t just busywork. It’s how I feel the collection. I want to see if the cards I love really hold up against the cards I think I love. That requires structure—but not rigidity.
| ID | Card | Rating | Matches | Rank | Tier |
|---|---|---|---|---|---|
| 13 | 2024 Aaron Judge - T210 | 1152 | 10 | 1 | S-Tier |
| 21 | 2022 Francisco Lindor - Topps Chrome - Blue Atomic - #95/100 | 1131 | 11 | 2 | S-Tier |
| 16 | 2019 Brandon Nimmo - Archives Snapshots - Auto , #03/50 | 1126 | 15 | 3 | S-Tier |
| 14 | 1978 Lenny Randle - Boxtop - Auto | 1125 | 15 | 4 | S-Tier |
| 6 | 2013 Carson Kelly - Bowman - Auto | 1105 | 17 | 5 | S-Tier |
| 25 | 2024 R.A. Dickey - T206 - Auto | 1095 | 6 | 6 | A-Tier |
| 20 | 2023 Pedro Martinez - 1988 Chrome Silver Pack - #67/99 | 1073 | 11 | 7 | A-Tier |
| 32 | 2011 Francisco Lindor - Bowman Chrome | 1061 | 6 | 8 | A-Tier |
| 17 | 2016 Aroldis Chapman - Gold Label | 1030 | 18 | 9 | B-Tier |