The market makers (index companies) take a view, or prediction, on what they feel will be the aggregate total of the results in any given market for all the races run.
Lets look at sps  (starting prices)
At Doncaster the on the 26^th of January 2001 the sp index was 55-59
The market makers believed that the winning horses sp, for each of the 7 races run at the meeting, would add up to around 57 points (the midpoint between the two quotes).
There were 98 runners at the meeting, and by looking at the statistical data for all jump races run on good ground with around 100 runners over 7 races, will show that the average make up (result) is in fact 46.9, so the index companies (in quoting 55-59) believe that the outsiders will do well at the meeting.
By using our data the user has to decide weather he feels the result will be above or below the index (the spread prediction by the index companies)
Since the average make up is 46.9 the logical conclusion is that the spread may be a little high, so the user may decide that the likely result will be less than the index of 55.
As with share trading, by selling the index the user would expect the market to fall or be lower than the index, and by buying the user would expect the market to rise or be higher than the index.
With the above scenario it would be a brave individual to expect the result to be above 59 since the average return is around 46.9.
Most people looking at this market and with the benefit of data available from spreadbets.org.uk would decide to sell this market, expecting the result to be less than 55. The next thing to decide is your stake.
Let us assume that you decided to sell the index at £1 per point, you then wait until all 7 races are run. The eventual result is 37; you sold at 55 so you win.
To work out your winnings you simply deduct the result from the price at which you sold (55). This is 18 so therefore you have won 18 times your stake (£1), which equals a £18 win.
Had you decided prior to racing that this was a fantastic money making opportunity and that the index was wrong, then perhaps you would have sold at £15 per point. that would have given you a profit equal to 18 times your stake, which would be £270 Profit.
But suppose you were unlucky and the outsiders did well and the eventual make up was 65, you would have lost 10 times your stake (the difference between your selling price 55 and the result 65)
This is in essence the basis of spread betting, after each race a new quote is given by the index firm to reflect the previous result. This gives the spread bettor the opportunity to either stay in the market and hope that results will improve in their favour, or get out (close) either taking profits, or limiting their loss. Generally speaking it is usually better in the long term, to stay in for the eventual result to determine profit or loss, rather than close during the meeting, since every time you trade you are paying trading costs, this is the gap between the buying price, and the selling price.
Should you decide to get out of your trade or "close" you would do so by simply trading in the opposite direction for the same amount of stake money. I.e. if you sold sps at 55, and after 4 races the quote is 37-40 , you would simply buy at 40 for the same amount that you originally sold for. Then regardless of the final result you have made 15 points (the difference between the selling price 55 and the buying price 40).