The In-Play market is totally different to the pre-race market.
The pre-race price shifts are due to changes in sentiment (its all about what's in people's minds). But the In-Play price shifts are due to actual physical changes in the position of the horses.
Indeed a very important distinction, and I agree with you 99%. By the end of the race, price shifts are indeed driven almost entirely by actual performance of the horses, but early on in races (especially longer races of 2m or more), there can still be several instances of price shifts due to sentiment rather than performance.
For example, I have encountered several cases where let's say I backed a horse at 20 and hoped to get out in play at 15 or 10. The horse comes out and seems to be running well, but nevertheless, someone has thrown up a large block of lay bets, at let's say 25. For a horse trading at 25 in a typical race, the amounts being traded in-play are usually traded in rather small amounts - 5 here, 10 there. If a horse is running well, but not "superbly", it could take quite some time for the market to eat through several hundred laid at 25.
In fact, usually by the time that has happened, the race has been decided - either the horse will fade and bidding to back the horse will disappear even before all of the lay bets at 25 have been absorbed OR the horse will be clearly challenging to win with odds or probably 5 or less. The point being, if you're backing-to-lay in order to secure a small/medium gain while being able to get out with only a small loss if the horse doesn't seem to perform early on, it's not possible - your back-to-lay has been essentially turned into almost a pure punt. I've had several cases where the horse performed well in the end, but I wasn't willing to risk the entire stake I was using for the race, so I had no choice but to get in front of the large layer and get out early with a small loss.
