Here.

If you go to the most recent paper, you can look at the tables, which compare different methods for dating, and by comparing numbers, see how the variation increases as you go deeper. 2000 years ago, the variations seem to be less than 200 years. As you get older, the times get less certain.

Accuracy depends on how far back you're going. Data for recent years also can pick up confirmation from other sources.

You can get some idea of the issue by looking at another variable, temperature, and seeing different methods match up.

Ice core data shows a definite relationship between temperature and CO2, but as yet, I have seen nothing definite that nails down the direction of initial causality (although there is a feedback issue later in the cycles.)