Being one who works in the gaming press, I have experienced just this problem. I often have to moderate or edit my team's writeups and articles to line up with publisher's expectations and to avoid receiving letters about reviews that are 'overly negative'. This often makes it difficult to give anything much below about 70% if you want to hear from the publisher ever again... we made that mistake with Codemasters and they're not very co-operative.

There are very few companies who will actually continue to co-operate with you after a negative review towards one of their products. I tend to find that reviewing products by dividing reviews up into two 'categories' of target audience, you achieve better results. By saying

This game is brilliant for those who don't care about the details of , but just want to kick back and relax with friends on a Friday night

you make it look good... and then simply follow it by

For those who are addicted to their games, you might want to wait until it hits the bargain bins.

That tends to mean you can average up your rating a bit... and no-one tends to notice - and your readers get to know your strategy after a while, resulting in 'accurate' reviews whilst keeping publishers happy.