On 20 October at 8.30am, Gordon Brown, the Chancellor, pressed the button that launched Sets, the Stock Exchange's new electronic system. It was a baptism of fire - the first few weeks of Sets' life coincided with some of the most volatile weeks the markets had seen since the crash of 1987.
From a technical point of view the system coped admirably, with barely a hitch in the two months since launch. But the Exchange failed to anticipate how traders would react to the new system, an oversight that has proved to be Sets' Achilles heel.
"We've created a sophisticated car, but people can't get it out of first gear," said Martin Wheatley, head of markets development at the Stock Exchange, and the man spearheading Sets.
Recent difficulties with Sets have prompted the Exchange to consider introducing changes in the new year - including shorter trading hours and a new formula for calculating the closing price.
The key problem with Sets is the discrepancy between how traders actually use Sets and how the Exchange would like traders to use it. This has caused numerous problems over recent months including countless trades executed at "rogue" - or unrepresentative - prices, numerous incidents of rogue closing prices for shares and a highly publicised instance of market abuse - as well as a number of less well publicised ones.
Fear of "rogue" or unrepresentative closing prices on New Year's Eve has prompted the Exchange to take the unprecedented step of intervening directly in market trading and disregarding closing prices deemed "exceptional" by statisticians. Closing prices on New Year's Eve are used as the basis for fund valuations, and so are particularly important.
Early morning and late afternoon illiquidity in the new electronic order book is the key factor behind the problem of "rogue" share prices.
Early in the morning, traders start to input their buy and sell orders into the book. The orders then remain on the book until one trader's "buy" order matches another "sell" order, at which point the trade is executed. But as many traders - particularly those with the larger institutions - do not start inputting orders until later in the day, there tend to be few orders placed on the book early in the morning, leading to so-called early morning illiquidity.
Late in the afternoon, traders tend to delete unexecuted orders from the book, so as not to get caught out by overnight developments in the Far East. So again, there are few orders on the book and the market is illiquid.
Generally, the greater the number of orders in the book, the smaller the difference between the lowest-priced "sell" order on the book and highest-priced "buy" order, known as the "spread".
First thing in the morning and last thing in the afternoon, market illiquidity means that spreads tend to be wide. At these times of the day, "sell" orders tend to be priced higher and "buy" orders tend priced lower than they would otherwise. So, if a trader decides to deal "at best" - that is to take the best price available - he runs the risk of dealing at prices that are unrepresentative of normal trading patterns. This can hurt unwitting investors and mean that a share's closing price - the price at which the last trade was executed - can be "rogue".
Not all "rogue" prices are accidental. Sometimes they result from deliberate manipulation, as was the case late in November when two JP Morgan traders tried to push down the level of the FTSE 100 index.
The two sold a number of bundles of pharmaceutical stocks "at best" in the late afternoon. The last bundle of stocks they sold was matched with a "buy" order that, because of market illiquidity, was priced substantially lower than one mighty ordinarily expect. This pushed down both the closing level of the individual stocks - SmithKline Beecham and Glaxo Wellcome - as well as the closing level of the FTSE 100. The traders subsequently lost their jobs and earned their employer, JP Morgan, a record pounds 350,000 fine from the Exchange.
The Exchange is likely to implement a variety of measures over the next year, but no radical reform is on the cards.
The most likely reform, which could be introduced in the first few months of next year, is early closure of the Stock Exchange and the introduction of a closing auction to determine the closing level both of the FTSE 100 index and of its constituent shares.
The form of the closing auction is yet to be decided, but one possibility is that the Exchange could allow no trades to be executed at a certain period of time at the end of the trading day - say the last 10 minutes. But orders could still be placed on the book, and the Exchange would run a computer program at the end of the 10 minutes which would match buy and sell orders and thus determine a stock's closing price.