With data on children’s dental state from 1976 to 1993, we evaluated how the size of the high-caries group has changed concurrently with simultaneously decreasing mean numbers of teeth with past or current caries. Information related to all dental check-ups done for 5- and 15-year-olds in Helsinki, some 4,000 subjects of each age by year, consisted of numbers of teeth with caries experience (dmft or DMFT) and of all decayed teeth (dt+DT). Polarization of dental caries was described as the proportion of high-caries groups in each year, both in terms of caries experience and current untreated caries, diagnosed at subjects’ annual clinical dental check-ups. For 5-year-olds, the high-caries group by caries experience included patients with their dmft ≥ 3. For 15-year-olds the limits were set at DMFT ≥ 6 and DMFT ≥ 15. The high-caries group in terms of untreated caries was similar for both age groups: dt+DT ≥ 3. Furthermore, polarization of caries was calculated as the share of numbers of both dmf or DMF teeth and dt+DT in each high-caries group of the total number of such teeth in the entire age cohort. During the 17 years, mean dmft for 5-year-olds decreased from 4.6 to 0.8 and mean dt+DT from 0.9 to 0.6. In 1993, 78% had their dmft = 0, whereas only 8% of the patients accounted for 76% of all decayed teeth. For 15-year-olds the decrease was even greater: their DMFT fell from 12.1 to 3.0 and their dt+DT from 3.1 to 0.8. However, only 26% had their DMFT = 0 in 1993, with 55% of all dt+DT occurring in 10% of the patients. The present results confirmed a strong polarization in caries for both age cohorts, showing the need for renewed strategies in preventive caries treatment.