Subject: Re: Market makers can be scammed
From: Peter C. McCluskey (pcm@rahul.net)
Date: Sat Mar 11 2000 - 16:12:31 PST
hallowel@math.unc.edu (Karl Evan Hallowell) writes:
>I'm not sure of the exact algorithm you're using here, but I'm disturbed
>that the market makers are throwing book (er... limit) orders that are a
>lot larger than anything else out there. E.g., in the Bush claim, I see 90
>and 72 as the size of the limit orders that the market maker has placed.
>Nobody else has committed to a larger order than 20. Suppose I had
>initially had the spread of the claim at 90-93 and everyone else was in on
>it. We could scam your market maker big time. True it would be much, much
>harder with larger numbers of traders, but I can see a tradition of
>fooling bots developing. IMHO, your bots should start with small limit
>order sizes and increase the order size as time progresses and more
>traders enter the market.
I guess you are right that there is a problem. I had thought about it,
and concluded that collusion would be no harder to detect than collusion
under normal trading, but neglected to think about the effects of the
larger rewards to cheating here.
I'm not enthusiastic about solving it by reducing the liquidity at
times when liquidity is needed most. I'm leaning towards a solution
that involves some human judgement before the claim is approved that
sets some limits on the initial prices of the market maker.
-- ------------------------------------------------------------------------ Peter McCluskey | Boycott Amazon.com until they stop suing http://www.rahul.net/pcm | companies that support 1-click shopping.
This archive was generated by hypermail 2b28critmail : Wed Aug 22 2001 - 05:25:00 PDT