The system works reasonably well, but it does not prevent systematic biases developing. This is impossible to avoid in a non proportional electoral system with single member constituencies.

One bias is caused by the Labour electorate being concentrated in mostly safe urban geographical areas; with a long term trend towards the electorate declining in the inner cities. This means that the longer it is since the last revision the smaller the average size of inner city seats and the larger suburban and rural electorates. This benefits Labour and disadvantages the Conservatives. Each revision therefore costs Labour a few seats purely for this reason (which is why the Labour majority in the 1966-70 Parliament voted to reject the results of a review; which was implemented by the Conservative majority in the next Parliament and came into force in 1974). However notice that the current English review is based on the 2000 electorate figures. Even when it first comes into force, it will already be a bit out of date with urban constituencies on average smaller than suburban and rural ones.

A related factor is that voters in safe seats, particularly Labour voters in safe Labour ones, are less likely to turn out to vote. This affects the aggregate vote statistics without costing Labour any seats.

Another bias, deliberately built in to the system, has been caused by Wales and Scotland being over-represented compared with England. Both countries are more Labour inclined than England. This bias has been reduced since the Scottish boundary changes in 2005, but not eliminated.

The Conservative Party has noticed the above two biases and has been doing some thinking. When they get the chance the Tories may introduce a system that aims more at mathematical equality of constituency electorates than the existing rules.

However the Conservatives do not seem to have recognised the other bias in the system, because to correct it they would have to adopt a more proportional election system.

The Labour Party vote is the most heavily concentrated, the Liberal Democrats vote is the most evenly spread and the Conservative vote is in the middle in terms of concentration.

You can see the effect by looking at the 1983 general election. The Conservatives got 397 of the seats in Great Britain for 43.5% of the vote, Labour (in its worst election result since before the First World War) won 209 seats for 28.3% of the vote, the Liberal/SDP Alliance received 23 seats for 26% of the vote.

To emphasise this the Labour Party even at a low level of support overall still won over 200 seats. The Conservative Party, in its bad loss in 1997, won fewer seats (165) for a higher percentage vote (31.5%).

As for the LibDems (and predecessors) you can see the systematic bias they labour under.

All this is due to relative concentration of support, combined with a single member electoral map.

Partisan gerrymandering is not really a factor, although Labour did supposedly have more influence in the last review by co-ordinating the submissions from their local Councils, parties and MPs (who sometimes have different priorities thus diluting overall party influence).

If you want an example of a donut constituency try York Inner and York Outer (replacing City of York and Vale of York), in the new distribution. However the Boundary Commission usually prefers to do an East/West sort of split.

As each review area is considered individually, different Deputy Commissioners approach similar problems in different ways.