Jigsaw Puzzle Association Rating (JPAR)

What Is JPAR?

The Jigsaw Puzzle Association Rating is a calculation of each member’s puzzling performance across USAJPA and sanctioned competitions that aims to:

• Capture your current skill level

• Reflect your improvement over time

• Accurately predict your future performance

• Indicate how your skills compare to the field of speed puzzlers

The higher your rating, the higher your speed and rank is in the speed puzzling field.

NEW! An Updated Algorithm

The October 2026 JPAR update (below) uses an entirely new algorithm than the previous updates. Though our previous calculated JPAR was a thoughtful approach, we wanted to build a system that more easily allowed for iterations over time, and corrected for some issues our keen volunteer data scientists noticed.

Our team regularly evaluates and iterates on the JPAR algorithm to flexibly account for key event variables and effectively achieve our aims. They dedicated a large part of 2026 to overhauling our JPAR algorithm and process. This new calculation now better reflects these goals when compared to historical JPAR results.

How Is JPAR Now Calculated?

For each event, we use competitors’ completion times and previous ratings to estimate puzzle difficulty and update their JPAR.

• First, we compile:

  • Completion Times for all USAJPA members who competed

    • If a member does not complete their puzzle within the maximum time limit, an estimated completion time will be calculated, which assumes the puzzler continues at the same average pace. 

  • Each participant’s previous skill estimate and the model’s uncertainty about that estimate

    • First time competitors begin with a common baseline and higher initial uncertainty.

• Then, we calculate:

  • Event Difficulty

    • Individual Performance: Each competitor’s time is converted to a logarithmic scale.

    • Average Event Performance: The model adjusts the log values based on competitors’ existing skill estimates. The values for all competing members are then averaged to estimate event difficulty

  • Confidence Factors: 

    • Each competitor’s result is compared with their previous skill estimate and adjusted for doubt, statistical anomalies, and volatility of performance by incorporating:

      • Uncertainty in the existing estimate

      • Expected variation in competition results

      • Time since the competitor’s previous event

    • A stronger-than-expected performance raises their new overall estimated skill, while a weaker-than-expected performance lowers it. You are truly competing against yourself.

• Then, we calculate the New JPAR.

  • Put everything together in an equation and you get your new JPA Rating!

New JPAR = e (1 - confidence) x previous JPAR + confidence x (log(time) - event difficulty)

How To Read The JPAR Spreadsheet

The JPAR spreadsheet has a handful of tabs:

• Events included (all the events that are included in the dataset)

• All member performance data

• Rating - sorted by name

• Rating - sorted by placement

  • Note: it may be exciting to see the number we call placement (100th in the country or 1,000th in the country), but please remember that this number is fairly volatile. Our community is growing and new members are added to this rating spreadsheet every quarter. Even if your placement drops, you still might be improving. We highly recommend focusing on your rating and less on your placement.

Current JPA-Ratings: October 2026