Run2 excel sumproduct

Skill cxcscmu/SkillLearnBench/skills/b2-self-feedback-claude-sonnet-4-6/weighted-gdp-calculation/run2_excel-sumproduct

[COLM'26] SkillLearnBench is the first benchmark for evaluating continual learning methods that automatically generate agent skills.

Install

npx -y skills add cxcscmu/SkillLearnBench --skill run2_excel-sumproduct

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

What its author says it does

Copied from the file, not written here

Calculate GDP-weighted mean of net exports as % of GDP using SUMPRODUCT in Excel, including all supporting statistics (min, max, median, percentiles).

SKILL.md

2.7 KB, as published. Nobody here has run it

Excel SUMPRODUCT: GDP-Weighted Mean and Statistics

Net Exports % of GDP (Step 2 Formula)

=ROUND((H12 - H19) / H26 * 100, 1)

H12 = Exports; H19 = Imports; H26 = GDP
Multiply by 100 to express as percentage points (e.g., 19.7 not 0.197)
ROUND to 1 decimal place per task requirements

Row Mapping in Task Sheet

Step 2 Row	Country	Exports Row	Imports Row	GDP Row
35	UAE	12	19	26
36	Bahrain	13	20	27
37	Kuwait	14	21	28
38	Qatar	15	22	29
39	Oman	16	23	30
40	Saudi Arabia	17	24	31

Statistics Formulas (H42:L47)

All statistics use the net exports % range for that year column:

Row 42 (min):     =ROUND(MIN(H35:H40), 1)
Row 43 (max):     =ROUND(MAX(H35:H40), 1)
Row 44 (median):  =ROUND(MEDIAN(H35:H40), 1)
Row 45 (mean):    =ROUND(AVERAGE(H35:H40), 1)
Row 46 (P25):     =ROUND(PERCENTILE(H35:H40, 0.25), 1)
Row 47 (P75):     =ROUND(PERCENTILE(H35:H40, 0.75), 1)

GDP-Weighted Mean (Step 3)

=ROUND(SUMPRODUCT(H35:H40, H26:H31) / SUM(H26:H31), 1)

Why This Works

H35:H40 = net exports % (already × 100)
H26:H31 = GDP weights
SUMPRODUCT(pct, GDP) / SUM(GDP) = weighted average
Since pct is already ×100, result is in percentage points

Mathematical Equivalence

Weighted Mean = Σ(pct_i × GDP_i) / Σ(GDP_i)
              = Σ((Exports_i - Imports_i) / GDP_i × 100 × GDP_i) / Σ(GDP_i)
              = Σ(Exports_i - Imports_i) × 100 / Σ(GDP_i)

Sample Values (2019)

UAE: 19.7%, weight 418.0B
Bahrain: 11.3%, weight 38.7B
Kuwait: 12.6%, weight 138.7B
Qatar: 14.3%, weight 176.4B
Oman: 12.5%, weight 88.1B
Saudi Arabia: 8.0%, weight 838.6B
Weighted Mean ≈ 12.2%

openpyxl Implementation

cols = ['H', 'I', 'J', 'K', 'L']

# Net exports % per country
for i in range(6):
    row = 35 + i
    for col in cols:
        task[f'{col}{row}'] = f'=ROUND(({col}{12+i}-{col}{19+i})/{col}{26+i}*100,1)'

# Statistics
for col in cols:
    task[f'{col}42'] = f'=ROUND(MIN({col}35:{col}40),1)'
    task[f'{col}43'] = f'=ROUND(MAX({col}35:{col}40),1)'
    task[f'{col}44'] = f'=ROUND(MEDIAN({col}35:{col}40),1)'
    task[f'{col}45'] = f'=ROUND(AVERAGE({col}35:{col}40),1)'
    task[f'{col}46'] = f'=ROUND(PERCENTILE({col}35:{col}40,0.25),1)'
    task[f'{col}47'] = f'=ROUND(PERCENTILE({col}35:{col}40,0.75),1)'

# Weighted mean
for col in cols:
    task[f'{col}50'] = f'=ROUND(SUMPRODUCT({col}35:{col}40,{col}26:{col}31)/SUM({col}26:{col}31),1)'