Triangular arbitrage

Discussion in 'Forex' started by bigspeculate, Feb 14, 2017.

  1. We know that Triangular arbitrage keeps a cross rate 'in line' with its respective majors.
    So that EUR/USD* USD/JPY = EUR/JPY.
    Is there any way to exploit this relationship for short term trading? Can one or two pairs lead the other? Can an indicator applied to one pair lead the others?
     
  2. toc

    toc

    Probably it is possible but should have already been exploited by the big players with supercomputers and powerfully complex algos.

    I would rather try to research on such arbitrage on longer time frames even within intraday like 30 min. instead of 1 min or few tick charts.
     
  3. truetype

    truetype

    No.
     
  4. Sig

    Sig

    Minor tip, a bunch of smart people make their living in trading. A subset of them look for every microsecond and fraction of a cent they can arbitrage and arbitrage the crap out of it until it goes away. So if it's obvious enough for someone who just learned about this whole trading thing last month to come up with, everyone else has figured it out too. In fact only one smart person in the entire universe has to have figured out a no risk arbitrage for it to go away. Or as truetype succinctly put it, no.
     
    SoesWasBetter and dealmaker like this.
  5. I never meant actual arbitrage, of course a retail trader can't trade for microseconds. Rather, can one use the movements of any of the above mentioned pairs as a leading signal for the related pairs? Something akin to intermarket correlations, or statistical arbitrage.
     
  6. toc

    toc

    Intermarket correlations are there for sure. A starting point would be 3 decades old book by John Murphy "Intermarket Technical Analysis". You can always research internet for such correlations between currencies, stocks, indicies, commodities.

    Also check how do Asian and European markets effect the others around the globe.

    Instead of trying to play the "millisecond" get ahead game, you might want to focus your efforts in catching a seat on the "bigplayers" rides i.e. major trends.
     
  7. quant1

    quant1

    Few points:

    - triangle arb is a "hard" arb. This means that it is truly risk free (provided one can get exposure while the mispricing persists). To get exposure, one needs to be very very fast. In my experience my best guess is the fastest players are shooting orders in single digit micros (maybe sub 1 micro). Milliseconds are an eternity compared to these times.

    - a soft arb is something akin to stat arb or pairs trading. These relationships are less stable and statistically risk free. If the correlation blows out you can lose money. These tend to be less latency sensitive and have lower alpha (due to higher variance)

    - cointegration and correlation exists on many time scales. However, the trade off tends to be variance as time increases. If you are interested in trading longer term stat arbs, transaction costs and modeling/cash management are key. I think equities have the most robust set to look for these relationships. Many also have options to create synthetics that have better leverage properties.
     
  8. Didn't Long Term Capital Management blow up when the correlations it traded went awry, during the Asian currency crises?
     
  9. Sig

    Sig

    Since any triangle arb would only last microseconds, how could it possibly provide you a leading indicator? I think you're throwing around a bunch of words without fully understanding the meaning behind them. It's very possible that there are statistical arbs between currencies. Those have nothing to do with the concept of triangle arbitrage any more than they do with the phases of the moon.
     
  10. Another example was the artificially created correlation where CHF was pegged to the EUR.
    I actually know a few guys that were playing this "sure thing" right up until the Swiss dropped the peg unexpectedly.
    The EURUSD/USDCHF/EURCHF triangle has been extensively explored on certain Forex-only web sites.
     
    #10     Feb 15, 2017