Or, here's an idea, maybe we should use a weighted index of several of those things as a measure, in order to prevent misleading data from the fluctuations in the price of any one item?
That's done already. More useful: such index as a ratio of wages. Say, [price of standardized basket of commodities] / minimum wage.
That would tell you how many # of paid work needed to (for example) fill a shopping cart with a week's worth of food + heating / powering one's home over that time.
("minimum wage" = applies to folks scraping by in low-paid jobs. Higher earners are living the good life anyway, imho).
Nah. Who would do that. We can just look at the price of gold, that's all that really matters.
See e.g https://www.macrotrends.net/1380/gold-to-oil-ratio-historica.... Over 80 years (since 1946), the ratio held stable within 10-40 range, with a few brief spikes. The variance from the average held in a band [-2x, 2x], with a few brief spikes of ~ +-4x.
See also e.g. https://www.creditdonkey.com/gas-price-history.html for the raw price of gasoline, exponentially shaped. The increase over 80 years (since 1946, to match the ratio dataset) is 15x.
Extra wrinkle. The exponent of the exponential is not quite constant over time, currency can debase much faster than one would naively expect by extrapolating historical trends.
Edit. For completness, the BLS inflation index tracks a price inflation from 1946 to 2023 of, guess what, 15x. https://www.in2013dollars.com/us/inflation/1946