The gaming of Oanda

Discussion in 'Forex Brokers' started by atticus, Mar 22, 2008.

  1. I didn't expect to uncover anything in a model-dependent environment, but I assumed if I found anything it would be the result of the American to European synthetic equivalence.

    A few weeks into the fun I began modeling neutral-delta binary range equivalence; in which a neutral touch/hit box is modeled with the minimum platform duration of five minutes. The "barrier/binary range" OTC exotic is a path-independent (Euro convention) exotic; price must strike[outstrike] a point inside[outside] two barrier strikes, but only at expiration. Any trade within those barriers will result in a payout at expiration on the short-gamma variety. The Oanda hit/touch synthetic offers a pure equivalent to the barrier range European exotic, in fact, the x-axis [5-minute duration] on the Oanda trade adds a small residual value over the European binary range.

    I found that the average OTC Euro barrier range hedge offered an 18bp edge over the Oanda "skinny hit" equivalent on EURUSD. The Oanda synthetic barrier range trade is actually superior on a touch due to the added duration. The Oanda position should always be priced at a small premium to the Euro barrier range.

    I began buying the Oanda hit/BR equivalent in size, but limited to $20,000 globally [user's SSN]. I began recruiting friends and family to open accounts with the proviso that they would not see the OTC hedge. I initially traded small-size to limit dealer-interest and began trading other major pairs to limit scrutiny from concentration in the EURUSD.

    It was cumbersome to trade OTC hedges in the Euro binary range, but the OTC vanilla straddle matched on duration was routinely paying double at barrier-touches. The vanilla straddle could also be gamma-traded at the outlier, but I also began to look internally with Oanda to hedge with American small delta hits/touches priced at the outliers. This was huge for the arb, but it required was the recruitment of traders for the long gamma hedges to circumvent the $20k global limit. The Oanda hedge required two independent touch purchases in lieu of the more common OTC trade which allows for a double barrier transaction. Recruiting Oanda accounts for hedging limited the scaling of the arbitrage, so we went back to the conventional straddle and barrier-range hedge with my primary dealer.

    Simply, Oanda was pricing the synthetic European barrier-range trade as an American variety -- adding as little as 5-minutes duration would decrease the payout 8-basis points.

    I am "knockout" on Oanda's forums.

    It's been discussed here:

    http://www.elitetrader.com/vb/showthread.php?threadid=120395

    Here is a screenshot showing the typical edge late in the game. The trade had already begun to be played-out. The betsfortraders.com price was a few bp higher than the indicative-price from interbank and superderivatives. If you look closely you'll see that the Oanda synthetic showed a duration exceeding an hour. I neglected to price the 5-minute duration for the screenshot, which would've brought the payout to $11.80.



    [​IMG]
     
  2. A delta-neutral American corridor [fwd start] option should not benefit greatly from a small increase in duration. It's essentially a value-added synthetic quasi-equivalent to the European range trade. Adding minutes in duration to the Oanda trade caused the payout to converge rapidly to the European range trade as modeled on superderivatives and what was seen as indicative pricing from dealers. Therein lies the pricing error. The arb was fat, but far more laborious to execute than the Oanda triangular arb from early 2007.

    Here is an example of the Euro range (path-independent) synthetic as traded on Oanda's platform. Priced as a "hit" box with a 5-minute duration:

    [​IMG]
     
  3. That pricing is downright 'silly.' The diction used makes this type of trade look overwhelming but if you take a step back and just look at it for what it is, it's algebraic. This isn't differential equations or Ito's Lemma. lol

    It's only a matter of time before the great Atticus finds another mispricing. I appreciate you threading this great find.

    I wouldn't be surprised to see you find something similar with CHF.
     
  4. lol, it's simply knowing what to look for.
     
  5. Very cool, bro. Thanks for the lengthy explanation. Your post gets pasted to a MS doc in my trading folder.
     
  6. Very true. Most people (myself included) are in a scattered conciousness and don't truly know what we are looking for in the market. All loss is the result of a scattered conciousness.

    Now, if you can find another arb that's scalable into the millions. :)
     
  7. "I didn't expect to uncover anything in a model-dependent environment, but I assumed if I found anything it would be the result of the American to European synthetic equivalence."

    My question is what is the process to which you were able to make this assumption?
     
  8. I assumed there may be a mis-pricing due to the fact that structuring is bimodal. Hits become misses depending on locale. I didn't expect to find it while trading against a quant-shop.

    That, and my mom always said I was special.
     
  9. Brother -- recall that you were offered a taste.
     
  10. That, I do recall. I cannot express my appreciation enough. Once I get all this legal shit behind me I'll be good to go, full steam ahead. I can't even get my S7 license because of it.

    The thing about me too is that I don't want to just be handed anything. I want to acquire the conciousness to be able to see it myself. I didn't have the luxury of having my mother tell me I was special, she was always too busy smoking and drinking and dodging punches from my dad.
     
    #10     Mar 23, 2008