Gonka Epochs 132-133 CPoC Compensation Analysis

Complete analysis and compensation calculations for participants affected by the Confirmation Proof of Compute (CPoC) bug in Gonka blockchain epochs 132-133.

Overview

This repository contains the complete statistical analysis proving that epochs 132-133 were abnormal, identification of affected participants, and fair compensation calculations based on median reward coefficients from successful participants.

Key Findings:

Epochs 132-133 were statistically abnormal: Confirmation ratios dropped 14-22% below baseline

45 affected participants: 13 in epoch 132, 32 in epoch 133

Total compensation: 24,799.48 GONKA

Selection criteria: CR 25-50%, miss rate Quick Start - Verify the Findings

Option 1: Review Final Results (Recommended)

Read the Summary:

cat COMPENSATION_SUMMARY.md

This provides the complete overview, methodology, and final compensation amounts.

Review Compensation Details:

JSON format: compensation_epoch_132.json and compensation_epoch_133.json

CSV format (Excel-friendly): compensation_epoch_132.csv and compensation_epoch_133.csv

Check Affected Participants Lists:

Epoch 132: affected_participants_epoch_132.json (13 participants)

Epoch 133: affected_participants_epoch_133.json (32 participants)

Option 2: Re-run the Entire Analysis

If you want to verify the calculations from scratch:

Install Python dependencies

pip install -r requirements.txt

Configure environment (first time only)

cp .env.example .env

Edit .env to set your archive node URL and inferenced binary path

Run analysis

Phase 1: Fetch epoch data and analyze confirmation ratio distributions

python3 fetch_epoch_data.py 132 133

Phase 2: Identify affected participants (those who received no rewards)

python3 identify_affected.py 132 133

Phase 3: Calculate compensation amounts

python3 calculate_compensation.py 132 133

Requirements:

Python 3.x

Python packages: pip install -r requirements.txt

.env file configured with:

ARCHIVE_NODE_URL (archive node endpoint)

INFERENCED_BINARY (default: /Users/fixtwin/gonka/gonka/inferenced)

Option 3: Spot-Check Individual Participants

Verify specific participants manually using the archive node:

0 → participant received rewards (not affected)"># Set up environment INFERENCED="/Users/fixtwin/gonka/gonka/inferenced" ARCHIVE_NODE="$ARCHIVE_NODE_URL" # From .env file

Example: Check if a participant received rewards in epoch 132

ADDRESS="gonka14ued4vcdeluj9v9vmsmteap7vtg7t50640hvmf"

Query their epoch performance summary (after settlement at height 2,060,000)

$INFERENCED query inference show-epoch-performance-summary-by-participant \ 132 $ADDRESS --node $ARCHIVE_NODE --height 2060000 -o json | jq '.'

Look for "rewarded_coins" field:

- If absent or "0" → participant received NO rewards (affected)

- If present with value > 0 → participant received rewards (not affected)

Key Files Overview

Analysis Scripts

File Purpose

fetch_epoch_data.py Phase 1: Fetch raw epoch data from archive node

identify_affected.py Phase 2: Filter participants by criteria (25-50% CR,

calculate_compensation.py Phase 3: Calculate compensation based on median coefficient

Output Files

File Description

COMPENSATION_SUMMARY.md START HERE - Complete analysis summary

compensation_epoch_*.json Detailed compensation records (ngonka amounts)

compensation_epoch_*.csv Same data in CSV format for spreadsheets

affected_participants_epoch_*.json Lists of qualified affected participants

analyzed_participants_epoch_*.json All candidates analyzed (including those rejected)

epoch_*_data.json Raw epoch data from Phase 1

Raw Data Files

File Description

epoch_132_data.json All 609 participants in epoch 132 with weights

epoch_133_data.json All 594 participants in epoch 133 with weights

Validation Checklist

Use this checklist to verify the analysis:

Verify affected participant criteria:

All have confirmation_ratio between 25% and 50%

All have miss_rate All have rewarded_coins = 0 (received no epoch rewards)

Verify compensation calculation:

Coefficient calculated from median of successful participants

Expected reward = confirmation_weight × coefficient

Compensation = expected_reward - actual_reward (0)

Spot-check sample participants:

Pick 3-5 random affected participants

Query their EpochPerformanceSummary from archive node

Confirm rewarded_coins = 0

Verify their confirmation_ratio is in 25-50% range

Verify totals:

Epoch 132: 13 participants, ~6,970 GONKA

Epoch 133: 32 participants, ~17,830 GONKA

Total: ~24,800 GONKA

Sample Verification Commands

Check a specific participant's qualification

Participant address

ADDR="gonka1hec2h63xkf9qf7gn07uwucveuhjfrqks8f4dmh"

1. Check their confirmation ratio (from epoch_132_data.json)

cat epoch_132_data.json | jq --arg addr "$ADDR" \ '.participants[] | select(.address == $addr) | {address, confirmation_ratio, base_weight, confirmation_weight}'

2. Check their performance stats (before epoch end)

$INFERENCED query inference show-participant $ADDR \ --node $ARCHIVE_NODE --height 2058356 -o json | \ jq '.participant.current_epoch_stats'

3. Check if they received rewards (after settlement)

$INFERENCED query inference show-epoch-performance-summary-by-participant \ 132 $ADDR --node $ARCHIVE_NODE --height 2060000 -o json | \ jq '.epochPerformanceSummary.rewarded_coins // "NO REWARDS"'

Verify the reward coefficient calculation

Sample a successful participant who DID receive rewards

SUCCESS_ADDR="gonka1w6wwv3wq25p8qge4lqsnfzs8lsd3s8ty6au65p"

Get their epoch summary

$INFERENCED query inference show-epoch-performance-summary-by-participant \ 132 $SUCCESS_ADDR --node $ARCHIVE_NODE --height 2060000 -o json | \ jq '{rewarded_coins, address}'

Get their confirmation_weight

cat epoch_132_data.json | jq --arg addr "$SUCCESS_ADDR" \ '.participants[] | select(.address == $addr) | {confirmation_weight}'

Calculate coefficient manually:

coefficient = rewarded_coins / confirmation_weight

Should equal ~52,796,590.93 for epoch 132

Questions or Issues?

If the verification reveals discrepancies:

Check archive node accessibility: Ensure the ARCHIVE_NODE_URL from .env is reachable

Verify block heights: Ensure querying at correct heights (see COMPENSATION_SUMMARY.md)

Check conversion factor: 1 GONKA = 1,000,000,000 ngonka (9 decimals)

Review selection criteria: Only participants with ALL three criteria qualify

Technical Details

Archive Node: Configured via environment variables (see .env.example)

Token Denomination: 1 GONKA = 1,000,000,000 ngonka (9 decimals)

Analysis Date: March 5, 2026

Epochs Analyzed: 132, 133

Total Affected Participants: 43 unique addresses (13 + 32 - 2 duplicates)

Recommended Compensation Pool: 24,799.48 GONKA

For questions or clarifications, refer to COMPENSATION_SUMMARY.md for detailed methodology.