[kijin]

To add a bit more data to your assertion, most of the 9,000 cases in South Korea can be traced to a handful of gatherings where people didn't wear masks.

First there was the cult in Daegu, where believers are not allowed to cover their faces because apparently their god can't see through fabric. Mass infections also occurred in churches where people took off their masks to eat or sing together, or tried to "sanitize" one another's mouths with a folk remedy. A gym in Cheonan was responsible for the majority of cases in that province, and recently there was a small burst of cases in an insurance company call center in Seoul.

Meanwhile, the subway in Seoul is just as packed with people as always, but I haven't heard of a single case that was conclusively traced to the subway. Why? Everyone wears a mask when riding the subway and sanitize their hands afterward.

Masks have been so effective that the Korean government has even stopped publishing locations that were visited by infected people if they are known to have worn a mask while visiting them. Yes, they still track down each and every case outside of the Daegu area.

[casperc]

Any chance you have some sources on this? There a lot of talking around the subject of masks, but I don't have any articles or similar to show people when discussing this.

[kijin]

The bit about masks is my own conclusion based on a lot of different local sources.

The closest thing to a comprehensive source about the distribution of infections in South Korea that I can find is [1], which still contains too much Korean-in-a-canvas that can't be automatically translated:

54.6% of cases are related to the Shincheonji cult

0.5+0.7+0.5% various other churches and religious groups

1.3% Daenam Hospital (almost all cases occurred in the psychiatric ward, as well as the attached funeral home where the Shincheonji cult leader held his brother's funeral)

1.2% related to gyms

1.2% related to the call center in Guro-gu

3.6% nursing homes

6.3% related to other mass infections (this includes other call centers, nursing homes, etc.)

11.3% contact with people infected from one of the other sources

3.3+0.6% infected overseas

14.9% other/unknown

Of course this doesn't say anything about masks. But you can clearly see that unlike in many other countries, ~70% of cases in Korea are concentrated in a handful of clearly delineated groups, with relatively little spillover into the general public. It takes a bit of familiarity with the religious landscape of Korea and the workplace culture to recognize that most of these groups are where people don't wear a mask. (As for nursing homes, the elderly often have difficulty breathing through a mask, not to mention they're together 24/7.)

Some people still got infected while wearing a mask, though, so it's not bulletproof. But it seems that the effect becomes visible as a kind of herd immunity when aggregated over a large population.

As for the government not publishing locations visited with a mask on, here's an example from my city's official Instagram account [2]. It says that the infected person's movements on some days are redacted because she wore a mask on those days (among other precautions) and therefore didn't make any epidemiologically significant contact with other people. You can compare this with other press releases by the same account that contain a lot more location data for other people.

[1] https://coronaboard.kr/en/#source-card-slide

[2] https://www.instagram.com/p/B-Iy3G7lTQs/