Spreads

'At touch' spreads

To calculate spreads we first replay order books and measure the best bid and offer prices every 30 seconds. The 'spread' is the difference between bid and offer prices divided by mid price. This is then converted to basis points (0.01%).

Where a book is empty or one sided or where the spread is very large (>200BPS), we mark that data value as 'bad' and assign a default spread of 201BPS for averaging purposes. If more than 10% of values are 'bad' for a venue then we exclude that venue from the results and show 'n/a'. The spread statistic we report for an instrument for a given venue is the average of all 30-second measured spreads over the period analysed.

The spread we report for an index is the average of the individual instrument spreads weighted by the traded volume of each instrument.

'Depth Weighted Spreads' - i.e. 25K EUR Spread

For a depth weighted spread, instead of measuring the simple best visible bid and offer in the order book, we 'match' a deal of a certain size up or down the best bids/offers available in the book until the deal size is fully matched.

So, for instance, if the best bid price in a book is 1.00 EUR with volume 20,000 and the next best bid is 0.99 EUR with volume 100,000 then the depth weighted 25,000EUR bid price is ((1.00*20,000)+(0.99*5,000)) / 25,000 = 0.998 EUR.

Apart from this initial step of finding depth weighted bids and offers, the methodology for the rest of the calculation of the depth weighted spreads is the same as for the 'Simple' spread.

Book Depths

To calculate depths we replay order books and measure the total value (i.e. price*volume) of all visible bids and offers within 'X' Basis Points of the mid price every 30 seconds. The value quoted for depth is then the average of these 30-second values over the period.

So, for example, if the mid price of a stock is 5.00 EUR, the '50BPS liquidity' figure will measure the total value of all bids and offers in the order book with prices between 4.975 and 5.025.

For indices the liquidity value quoted is the average of the depths of all the constituents of the index.

Best Prices

'At touch'

For this statistic we simultaneously replay all order books for an (ISIN, CCY) instrument for all venues on which it trades. Then, every 5 seconds we look at the best bid prices on all venues and see if there is either a single 'winning venue' with the absolute best bid price or a tie. Once this is done for a complete trading day we record the percentage of time each venue has the absolute best price. Although we measure this separately for bids and offers, we generally average the results over the bids and offers when reporting.

These percentages do not add up to 100%! For instance, say the percentages are (LSE 20%, CHIX 20%, TRQX 10%, BATE 8%, NURO 5%). This would imply that (100-(20+20+10+8+5)) = 37% of the time there was a 'tie' for the best price in the market and no one venue was 'best'. The other 63% of the time a single venue had the best price.

'Depth Weighted Best Prices'

For depth weighted best prices, instead of using the best bid or offer price in the book, we 'match' up or down the book to get weighted bid and offer prices for different size deals [same methodology as for depth weighted spreads, see above]. Apart from this way of determining the best bid and offer prices, the rest of the methodology for determining the At Best Percentages is as before.

Depth weighted Best Price percentage statistics, especially for larger deal sizes, yield fewer ties between venues. This is because even if two or more venues share the same best bid/offer price, using depth weighted measures means that the volumes available at the best price and also the prices and volumes below the best price come into play.

The spread we report for an index is the average of the individual instrument spreads weighted by the traded volume of each instrument.