Their impact will be far less than might appear

The two pollsters who are carrying out regular London Mayorals surveys, and ComRes and YouGov are both issuing two sets of numbers when they publish their results. These are the first preferences of those interviewed and then their estimation of what the final split between the top 2 will be.

The media is tending to focus on the latter which might be misleading.

My view is that we should put our main emphasis on on the former.

The experience of all the mayoral elections that have taken place in the dozens or so authorities where there are elected mayors is that the “supplementary vote”, as it is known, rarely affects the final outcome.

There have only been two exceptions, at North Tyneside in 2005 and at Doncaster, where the winner was not the first in the first round. There are several reasons.

Firstly a surprisingly large proportion of voters don’t exercise their supplementary votes being content just to exercise a single choice

Secondly the 2nd preferences of those who go, say, Ken-Boris or Boris-Ken, don’t count

Thirdly those whose 2nd preferences are not for one of the top two on the first round do not have an impact on the outcome.

This can be quite hard to poll.

My guess is that the valid second preferences will split broadly in line with the first preferences

@MikeSmithsonOGH