Comments on: In a blow to the French, BetFair choose Bastille Day to premiere the revised version of the bet-matching logic of their prediction markets. — IMPROVEMENT MEANS BETTER LIQUIDITY FOR THEIR EVENT DERIVATIVE TRADERS. http://www.midasoracle.org/2008/07/14/betfair-bet-matching-logic-7/ Prediction Markets For All Thu, 20 Nov 2008 08:24:26 +0000 http://wordpress.org/?v=2.6.3 By: Ed Murray http://www.midasoracle.org/2008/07/14/betfair-bet-matching-logic-7/#comment-20534 Ed Murray Mon, 14 Jul 2008 15:28:29 +0000 http://www.midasoracle.org/?p=7508#comment-20534 <strong>Pre-2008</strong> - $50 is available in the backing column at 1.8 (and another $40 is available to lay at 2.25 on the opposite player), and Chris F. Masse submits a $100 bet at 1.8 backing player A, $50 is available and matched at 1.8, the other $50 remains up on the board at the price he asked for, going into the lay column.  Then Ed Murray 30 seconds later submits a lay bet of $50 at 1.8, and the remainder is matched.   However, there was 2.25 available to lay on the opposite tennis player, which could have matched Chris's bet 30 seconds earlier, but was left up on the board in a 100% book.  - <strong>2008 new bet matching logic</strong> - $50 is available in the backing column at 1.8 (and another $40 is available to lay at 2.25 on the opposite player), and Chris F. Masse submits a $100 bet at 1.8 backing player A, $50 is available and matched at 1.8, the other $50 is also matched against the $40 lay at 2.25 of player B.  Chris's bet is fully matched and therefore there is none of Chris's money left up on the board.  Then Ed Murray 30 seconds later submits a lay bet of $50 at 1.8, and the bet stays up on the screen.  - - Overall there is no improvement in liquidity, Chris gets his initial bet matched, but it also becomes harder for subsequent players to match Chris's money.  The benefit of a quick match for Chris is exactly cancelled out by the difficulty of a future bettor trying to match Chris's money, which becomes much harder.  - Conclusion: Overall liquidity is completely unaffected by the new bet matching logic. Pre-2008
-
$50 is available in the backing column at 1.8 (and another $40 is available to lay at 2.25 on the opposite player), and Chris F. Masse submits a $100 bet at 1.8 backing player A, $50 is available and matched at 1.8, the other $50 remains up on the board at the price he asked for, going into the lay column.  Then Ed Murray 30 seconds later submits a lay bet of $50 at 1.8, and the remainder is matched.   However, there was 2.25 available to lay on the opposite tennis player, which could have matched Chris’s bet 30 seconds earlier, but was left up on the board in a 100% book. 
-
2008 new bet matching logic
-
$50 is available in the backing column at 1.8 (and another $40 is available to lay at 2.25 on the opposite player), and Chris F. Masse submits a $100 bet at 1.8 backing player A, $50 is available and matched at 1.8, the other $50 is also matched against the $40 lay at 2.25 of player B.  Chris’s bet is fully matched and therefore there is none of Chris’s money left up on the board.  Then Ed Murray 30 seconds later submits a lay bet of $50 at 1.8, and the bet stays up on the screen. 
-
-
Overall there is no improvement in liquidity, Chris gets his initial bet matched, but it also becomes harder for subsequent players to match Chris’s money.  The benefit of a quick match for Chris is exactly cancelled out by the difficulty of a future bettor trying to match Chris’s money, which becomes much harder. 
-
Conclusion: Overall liquidity is completely unaffected by the new bet matching logic.

]]> By: Ed Murray http://www.midasoracle.org/2008/07/14/betfair-bet-matching-logic-7/#comment-20532 Ed Murray Mon, 14 Jul 2008 15:13:26 +0000 http://www.midasoracle.org/?p=7508#comment-20532 There are two advantages from this - (1) - When you submit a bet it looks for the best price for you across more than one selection, rather than just the single selection (2) - Betfair look to have changed towards a much healthier PR strategy, with a much more positive reaction from their users, and have stopped the "spin" tactics, which annoyed pretty much every Betfair customer who posted comments about the logic put up in spring 2008. - There is one disadvantage - (1) Money is still being taken risk-free into Betfair coffers by matching people at best increment and not best price (there is a significant if small difference between giving someone 1.83 and 1.8333333 say).  Its still free money for Betfair, and will have a negative effect on liquidity (all commission/rake/skims on a market will decrease the amount wagered over time on the markets - the higher the cost of placing a bet on Betfair, the less bets will be made, so at the margin, because of the small risk-free skimmed money, liquidity will fall because of this bet matching logic). - - There are plus points to the latest bet matching logic.  Overall I think it is a good thing.  However, extra liquidity clearly isn't one of them.  Extra liquidity would come from more users betting as a result of the new bet matching logic (why would this logic in itself attract any extra users to the site?), or from the same current users betting more (which is unlikely to happen, as Betfair are still incorporating a skimming feature giving them extra profits on top of their commission).  - The real effect is a temporal one.  If you submit a bet there might be £1000 available at the price you ask for, but with the new logic, BF will match your bet laying other selections as well as backing the one you for example wanted to back.  Therefore your bet will be matched quicker, however, less bets will then get left up on the exchange.  The increasing attractiveness of submitting a bet (you are more likely to get matched instantly) is offset by the increasing unattractiveness of leaving a bet up, relatively speaking (which has to fall relative to taking a price off screen, if taking the price becomes more attractive).  It works in the same way as perhaps a credit card does - you get the chance to buy things now, with money you haven't yet earned, swapping consumption in the future, for consumption now.  The real effect isn't an increase in liquidity on the Betfair markets, its solely easier to get on straight away, and the overall liquidity on an event will stay the same (or fall slightly if the commission takes a bite at the margin, which it will do to a fractional extent).  - I'm in favour of the new bet matching logic, but linking it to an increase in liquidity is a red herring.  The total liquidity on each event will probably fall overall, as a result of the non-best execution part of the new logic.  Perhaps it is like supermarkets delivering to your door - you will get your food much more quickly, but it is unlikely you will actually want any more food.  If you want to back tennis player A, that is the same as laying player B, and the logic looks for best price on both ways of wagering on player A to win a match.  Overall liquidity is linked to the cost of using Betfair for Betfair users, on supply and demand curves, and this bet matching logic makes it easier for someone to get matched submitting a bet, BUT that bet is less likely to be left up on the exchange as a result, for other people to take, so the effect on liquidity is completely neutral, ceteris paribus, and assuming zero Betfair skimming profits. There are two advantages from this
-
(1) - When you submit a bet it looks for the best price for you across more than one selection, rather than just the single selection
(2) - Betfair look to have changed towards a much healthier PR strategy, with a much more positive reaction from their users, and have stopped the “spin” tactics, which annoyed pretty much every Betfair customer who posted comments about the logic put up in spring 2008.

-
There is one disadvantage
-
(1) Money is still being taken risk-free into Betfair coffers by matching people at best increment and not best price (there is a significant if small difference between giving someone 1.83 and 1.8333333 say).  Its still free money for Betfair, and will have a negative effect on liquidity (all commission/rake/skims on a market will decrease the amount wagered over time on the markets - the higher the cost of placing a bet on Betfair, the less bets will be made, so at the margin, because of the small risk-free skimmed money, liquidity will fall because of this bet matching logic).
-
-
There are plus points to the latest bet matching logic.  Overall I think it is a good thing.  However, extra liquidity clearly isn’t one of them.  Extra liquidity would come from more users betting as a result of the new bet matching logic (why would this logic in itself attract any extra users to the site?), or from the same current users betting more (which is unlikely to happen, as Betfair are still incorporating a skimming feature giving them extra profits on top of their commission). 
-
The real effect is a temporal one.  If you submit a bet there might be £1000 available at the price you ask for, but with the new logic, BF will match your bet laying other selections as well as backing the one you for example wanted to back.  Therefore your bet will be matched quicker, however, less bets will then get left up on the exchange.  The increasing attractiveness of submitting a bet (you are more likely to get matched instantly) is offset by the increasing unattractiveness of leaving a bet up, relatively speaking (which has to fall relative to taking a price off screen, if taking the price becomes more attractive).  It works in the same way as perhaps a credit card does - you get the chance to buy things now, with money you haven’t yet earned, swapping consumption in the future, for consumption now.  The real effect isn’t an increase in liquidity on the Betfair markets, its solely easier to get on straight away, and the overall liquidity on an event will stay the same (or fall slightly if the commission takes a bite at the margin, which it will do to a fractional extent). 
-
I’m in favour of the new bet matching logic, but linking it to an increase in liquidity is a red herring.  The total liquidity on each event will probably fall overall, as a result of the non-best execution part of the new logic.  Perhaps it is like supermarkets delivering to your door - you will get your food much more quickly, but it is unlikely you will actually want any more food.  If you want to back tennis player A, that is the same as laying player B, and the logic looks for best price on both ways of wagering on player A to win a match.  Overall liquidity is linked to the cost of using Betfair for Betfair users, on supply and demand curves, and this bet matching logic makes it easier for someone to get matched submitting a bet, BUT that bet is less likely to be left up on the exchange as a result, for other people to take, so the effect on liquidity is completely neutral, ceteris paribus, and assuming zero Betfair skimming profits.

]]> By: Is the BetFair’s brand-new bet-matching logic (which matches bets across related selections) the first time a prediction exchange manages to increase liquidity WITHOUT augmenting the number of traders or relying on an automated market maker? | Midas http://www.midasoracle.org/2008/07/14/betfair-bet-matching-logic-7/#comment-20529 Is the BetFair’s brand-new bet-matching logic (which matches bets across related selections) the first time a prediction exchange manages to increase liquidity WITHOUT augmenting the number of traders or relying on an automated market maker? | Midas Mon, 14 Jul 2008 13:26:59 +0000 http://www.midasoracle.org/?p=7508#comment-20529 [...] Site Map In a blow to the French, BetFair choose Bastille Day to premiere the revised version of the bet-matc... [...] [...] Site Map In a blow to the French, BetFair choose Bastille Day to premiere the revised version of the bet-matc… [...]

]]>