The reason Bolsonaro's vote share will ultimately fall, can be explained relatively simply: The southern states that are more right-leaning are currently counting faster than the northern states, that are more left-leaning.

If I had to say, I don't think the outcome will be a repeat election, but I really don't know. Either way, it is a serious situation that necessitates changes to combat future attempts at election fraud.

It's unclear what happens now, and I will not claim to be an expert on it. Theoretically, a repeat election across the entire constituency is possible and is being talked about in parts of the media. If that were to happen, the bloc majority in the Riksdag could change.

It should be emphasized that this is still being investigated and does not concern the counting of votes itself. The result in the municipality also could not have moved candidates between the parties. But it might have influenced what candidates within the parties got elected.

Update: SVT is now projecting, as has been kind of apparent for a while, that the blue bloc will not win a majority of seats.

At this point, I am fairly confident that V ultimately will win the 349th seat. So the total bloc distribution will be 176 to 173 seats.

So bottom line is this: The late-arriving vote count is going approximately as we would have expected and the overall bloc majority will not change.

The reason is that while gaining the 349th most marginal seat (right now held by S) requires relatively few votes, the blue bloc doesn't just need to gain the 349th seat, but also the 348th seat (held by MP). And there more votes are needed.

It is highly unlikely that the bloc lead will change. I have seen some people incorrectly think that some 1,500 votes for M and SD each is enough for both to gain a seat. This is not correct and is a misunderstanding of the math. You need substantially more votes for that.

Inkscape

The only believable path to 175 seats I can find, is if in contrary to expectations, M overperforms by an extraordinary large margin on these votes, so much so that it reaches 72 seats and flips a seat from both S and MP. Technically possible, but would be surprising.

You can see that these votes slightly favor M. And this is why I expect M to gain a seat from S. As happened in 2022 with that year's late votes.
However, unless the tendency is in the opposite direction of previous patterns, it's difficult to see that last seat flipping.

In hindsight I could have tried to also include regional breakdowns in my projection. Although looking at the numbers right now, I am not sure if it would have fully captured the geographic shift.

For the exit poll the issue is mostly the margin of error. The miss for most parties was acceptable. But for M and MP, i was larger than you want it to be. One possible explanation that has been mentioned, is that this could be caused due to the sample of young voters not being fully representative.

For the polling miss, I think it's arguably not even a miss, as much as it was just the fact that the decisive voter movement towards M happened in the very last few days of the election, whereas most polls were published on Friday, two days before the election.

That's probably not the cause of the polling miss. The reason my own projection was like that is because it's based on the 2022 election result, not taking into account relative geographic swings.

For the blue bloc to win 175 seats, SD would need to win a seat from MP. While not completely impossible (there is theoretically a path), this seems highly unlikely to me. You have to consider that in 2022 SD underperformed nad MP overperformed among those votes.

Based on the result in 2022, I expect both M and V to gain ever so slightly, while S declines ever so slightly. I think M wins the last seat back from S, but I am far from sure.

For those curious, the assessment I had on election night is essentially unchanged. This is what the current seat calculation looks like (without trying to project remaining votes). It is very open what party wins the last seat, but it's unlikely the blue bloc wins a majority.

As it states in the image, this is the result with all regular precincts counted, but without the late-arriving ballots, which will be counted on Wednesday.