From: the.wise.one@abel.co.uk   
      
   Failing the model test   
      
      
   My first real taste of model failure was spectacular.  I had had a few   
   blow-ups before but they had been manageable, I had kept my job.  Then,   
   I did a six month currency option for $1,000 million.  It was for a very   
   good client.  The sales desk put lots of pressure on me: I foolishly   
   obliged.  It was an Australian dollar (A$) put/US dollar (US$) call with   
   a strike price of A$1:US$O.7210.  I was giving the client the right to   
   buy US$1,000,000,000 from me.  In return, they would give me   
   A$1,386,962,552.  They would only exercise the option if the A$ fell   
   below US$0.7210, say to US$0.70.   
      
        I traded in the currency markets to hedge the option.  This meant   
   that I sold A$/bought US$ according to our pricing model.  The idea is   
   that you match the delta of the option with the delta of the currency   
   trade.  If the A$ was weaker and the option was likely to be exercised,   
   then I sold more A$/buying more US$.  This was to offset the A$ that I   
   would receive when the option was exercised and the US$ I would have to   
   hand over.  If the A$ went up then I reduced my hedge as the option was   
   less likely to be exercised: this is called delta hedging or dynamic   
   replication.  It is supposed to be so simple that a monkey can do it.   
   That's why I got the job.   
      
        Things went fine.  Finally, it was the day before expiry.  The   
   exchange rate was A$1:US$0.7330.  The option was out-of-the-money -- It   
   had no value.  The delta was close to zero.  The model said I did not   
   need to hedge.  The option would expire unexercised.   
      
        The option was due to be exercised at 3 pm Tokyo time; the trading   
   day opened; I was at the usual morning meeting on the trading floor.   
   Suddenly, I heard a lot of screaming on the broker lines -- someone, a   
   hedge fund it turned out, was aggressively selling the Aussie, A$'s.   
   The hedge fund traders had used the illiquid, dozy early morning Sydney   
   market to sell a large amount.  The A$/US$ exchange rate plummeted to   
   around A$1:US$0.7220.  Other traders were also selling now.  By   
   mid-morning, the exchange rate settled around A$1:US$0.7208/0.7215.  My   
   bowels were anything but settled.   
      
        The exchange rate was now around the strike of the option, there   
   was a strong chance that the option may be exercised.  The delta of the   
   option was now close to 0.50, it was now a 50:50 bet.  I had been unable   
   to hedge as the exchange rate fell suddenly; the market had gapped; I   
   was completely unhedged, stark naked.  If the option was exercised, each   
   0.0001 (1 FX point) would result in a loss of around US$139,000   
   (A$192,000). I had been 'gamma-ed'.  This is the risk that option   
   traders fear most.  It gets every option trader in the end if you sell   
   options.   
      
        I was in a dreadful bind.  I had to decide whether the option was   
   going to be exercised (spot below US$0.7210) or not exercised (spot   
   above Us$0.7210).  If the former then I should buy US$1,000 million and   
   sell the equivalent A$.  This would exactly cover the option cash flows.   
     If the latter, then I shouldn't hedge.  The only problem was that I   
   didn't have any idea what was going to happen next.   
      
        I was learning who my friends were.  I seemed to have developed a   
   bad case of halitosis judging by the wide berth that everyone was now   
   giving me; they sensed an imminent death.  The risk manager helpfully   
   advised me that I was in breach of my trading limits; they would have to   
   report me.  The quants couldn't believe that the market had gapped so   
   much.  The trading room was run by an ex-soldier who had never traded.   
   He was there to provide 'management' and 'leadership'; he had boasted of   
   his combat experience.  He was in an absolute panic, alternating between   
   abusing me and throwing up.  I was contemplating life after finance,   
   perhaps driving cabs.   
      
        Marty saved me; he was the chief currency trader; we had a mutual   
   love of cricket.  He told me not to try to hedge the option's delta.   
   'Let's see what turns up.'  It was Mr Micawber's Dickensian trading   
   model.  We monitored the position closely, or as closely as my frequent   
   trips to the toilet allowed.   
      
        Marty figured that the traders at the hedge fund had taken an   
   opportunistic, short-term position; they would look to take profits.  He   
   worked this out via a series of seemingly aimless desultory   
   conversations with other dealers and brokers.  It was time for /The Art   
   of War/: 'Draw them in with the prospect of gain, take them by confusion.'   
      
        Marty waited till the Tokyo dealers were coming back from their   
   traditional noodle lunch.  He bought a modest amount of A$, tactical   
   buying; he traded casually through inter-bank brokers and with other   
   dealers directly; he was giving the market the idea that there was   
   interest in picking up the Aussie at these levels.  The exchange rate   
   drifted up to around A$1:US$0.7235/0.7242.   
      
        It was time for 'shock and awe'.  Marty now instructed his traders   
   to initiate a sudden and sustained bout of A$ buying, hoping to provide   
   the impression of a large buy order.  I had undergone a brief religious   
   conversion and was praying.  The A$ gapped up.  The hedge fund was   
   suddenly looking at losing the substantial profits on the trade; they   
   joined the frenetic buying; they were buying back the A$'s they had sold   
   earlier.  Marty was selling the A$s purchased moments earlier, probably   
   to the hedge fund as it was trying to turn its position around.  We were   
   out of jail.   
      
        The exchange rate ended the day at US$0.7308/0.7315, little changed   
   from its previous day's close or morning open.  The option was not   
   exercised and we had actually made money from the currency trading.  It   
   had been about six hours.  I had aged about 15 years.   
      
        The head of trading rang London to tell them of his personal   
   brilliant leadership and management that had averted a potential   
   disaster.  Marty told me that our fearless leader had never seen action   
   -- for my part I had seen enough.  I learned enough to never again run   
   such a large position and never ever be net short options on this scale.   
     I couldn't take the gamma.  Others had different ideas.  Their stories   
   had indifferent endings.   
      
      
   from:   
   "Traders, Guns & Money: Knowns and unknowns in the dazzling world of   
   derivatives"   
   by Satyajit Das   
   FT Prentice Hall, 2006   
   Chapter 6: Super Models -- derivative algorithms   
   pages 200-202   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)