I agree but if it was 1 then I would be very interested in the details.
So would I.
There was a news item the other day which reported that England counts its coronavirus deaths differently to the other home countries. The official statistic is the number of people who have died in care homes and hospitals having previously tested positive for coronavirus. Scotland, Wales and Northern Ireland were applying a month limit to how long previously the test could have happened. i.e. if somebody tested positive in March and died in May, in England it was a COVID19 death, elsewhere it was not.
If England had decided to retroactively bring its counting method in line with Scotland etc, you would see an apparent reduction of deaths in England. I'm not saying it is that specific thing, but some other similar counting anomaly might be the cause.
I've done a bit more digging. The graphics source is the European CDC which is the same source used by ourworldindata.com. Check out this graph of UK daily and cumulative cases:
https://ourworldindata.org/grapher/total-and-daily-cases-covid-19?country=~GBROn 3rd July, there was a massive step down and approximately -30,000 new cases. Obviously you can't have a negative number of new cases, so somebody did an adjustment on that day. The same step down appears in data direct from the government.
https://coronavstats.co.ukSo I think option 1 is the most likely, after all.