HiI’m currently trying to implement a replication barrier pricing model based on Wystup article http://www.wystup.com/papers/OT_derivativesweek.pdf. This method is commonly used on the FX market, but still, I’m facing some difficulties to match market prices – especially on emerging markets :*/ In some case, it seems that I should not take into account the butterfly in my hedge basket : does anyone use different hedge portfolio for different barrier types / markets ? Should I always use the most liquid delta or the best ratio sensitivity/price ?*/ I’m currently weighting my portfolio by the probability of not touching the barrier. When the barrier is too close, my price adjustment is almost null, whereas the market price is far off the TV… any idea ?Thanks a lot for your help… I’m not given so much help on the desk

Hi, I assume you are a FX quant. Most banks use their own proprietary method to compute this "fudge" factor, which kind of explain why no one is responding to your mail Weighing your replicating portfolio by the no touch prob will not get you anywhere close to the market except for certain cases of knockouts (not the reverse kind) but even then, it does not work well for all knockout cases. The earliest users simply use a factor like 0.5 over a range of one touch prob for options with parity, whch still seem to work reasonably for low skew ccy pairs like EUR/USD. The challenge is to make it work for USD/JPY. I believe most banks are moving on to stochastic volatility models but getting an accurate heuristic model working is worth the effort as you would want to calibrate your stoch vol model to market barrier prices. The latter can be obtained with some confidence from your heuristic model without the need to consult a broker.