Hi All,
New to the forum and I thought I’d post my experiences so far and hopefully help a few people, and conversely get some info on what others are doing.
I started currency trading in 08 when I sold my business (Australia) to a US company. Realising how much currency movements were going to affect my final sale amount I locked in the sale at 63 US cents (per 1 AUD). This turned out to be a masterstroke and by the time the money came through the rate was substantially higher than this and I would have seen a lot less cash.
This got me started and after trying quite a few platforms I went with Oanda as they seemed to be the easiest to use at the time. After a few ups and downs and some emotionally panic’d trading I started spending more time on fundamentals, thankfully at the time it was fairly obvious that the US dollar was extremely overvalued and I got into carry trading. Buying 500k of TRY and HUF against the USD in 2009, this not only saw a massive daily interest gain (over $250 per day for some time), but also a great profit after about 9 months.
This taught me to really read up on fundamentals and particularly when a currency is exceptionally overvalued or undervalued look at what you can do for longer term trades. However, for these types of trades I always use daily graphs, wait until a turn has started to occur (don’t guess what you think the top will be) and of course set a stop that you can handle losing.
After that I decided to investigate robot trading, with a self-taught programming background I grabbed eSignal (as their programming is all very easy javascript) and set to work writing formulas and such. eSignal doesn’t get their data from Oanda so the rates can be slightly different, but not enough to be a problem unless your scalping. They have a plugin however so you can enter market and limit trades. 60 min charts can backtest for about 3 years, 5 min charts backtest for just over 5 months.
Obviously formulas cannot take into account any fundamentals. The first formulas I found to be quite useful were using 60 min graphs and I would pick just after a pivot point turnaround, however, over the last couple of years there has been a lot of trending movement and these weren’t overly successful. I never implemented any of those live. I then moved to pivots but only on trades where trending was in place in that direction. This was more successful but alone still not profitable. I added in some money management to lower my average (similar to buying allocations of shares), and came up with formulas that worked on most currency pairs, however, in some cases the drawdowns could be fairly dramatic. Regardless I ran these for about 12 months and made some good money overall on them.
Mid 2010 however, I had a few failures on these (playing with reasonable amounts of money each failure costs $15-25k), and decided that the risk level was higher than it needed to be, with current volatility I would head back to fundamentals for a while.
Well, the problem with that of course is that I had been letting formulas go without worrying about any emotional trading. It’s hard to take a loss and I have a couple of short AU trades that’s are down about 40k right now…ouch. Obviously I wasn’t watching the fundamentals closely enough when I put those in and made a common newbie mistake of letting my losses run. Still confident it will come back as AU is overvalued medium term, but short term there may still be further growth, but that ties up some margin at the moment and of course I would have been far better to take the loss initially and not get emotional about it.
So now I’ve implemented a new formula that deals with breakouts, has short stops and take profit points and some probability based money management. Backtested on three currencies I’m seeing excellent results (57-67% profitable trades and profits are always higher than stops, with a max drawdown about 1/5th of profit). As this is only 5 minute charts I can only backtest 5 months. So far so good on this.
Is there anyone around that focuses on formulas rather than fundamentals, and if so what kinds of results do they get when backtesting their formulas? I’ve be very interested to know.
Cheers,
- Will