Tuesday, August 31, 2010

Man On Wire - tonight!

Just a quick note for those in Ireland/UK - the epic Man on Wire documentary (probably my favourite film of the last few years) is on Discovery Channel tonight, 9pm.

If you haven't seen it - don't miss!

Friday, August 13, 2010

Lupe Fiasco - "Dumb It Down"

I don't listen to a huge load of hip hop - an occasional work trip to LA means that a convertible and cruising the freeways to some chord-laden West-coast rhythms somehow suits the SoCal vibe - but when you really want some serious rap to make your brain stop and back up, Lupe Fiasco has got to be top of the game.

Try "Dumb It Down" from a few years back: an utterly hypnotic beat that keeps stepping up and down just by swapping the kick drum sound, a deathly dark synth line, and outrageously conscious lyrics implicitly lancing gansta rap and everything that goes with it.

You been sheddin' too much light Lu - you make them wanna do right Lu
They're getting self-esteem Lu - these girls are trying to be queens Lu
They're trying to graduate from school Lu - they're startin to think that smart is cool Lu
They're trying to get up out the hood Lu - I'll tell you what you should do...

Monday, August 02, 2010

Buy buy sell sell!

I've recently been having a gander at stockmarkets and the like. There's something that I find utterly fascinating about it - it's clearly a casino and a place to lose a stack of money.

But more and more I can't help but wonder, is there any way that you could take a little slice of money out of it, if you were prepared to trade really quickly, pay huge amounts of attention, and weren't hungry for big profits?

I've been taking a look at the trading options offered by GFC Markets (you've probably seen the banner adverts all over the internet - profit from the falling Euro!). Ignoring the gold, oil, and normal stocks that you can do through them, one of the main things they offer is trading in currency pairs (i.e. buying euros against dollars, or selling Swiss Francs against the Yen - whatever). Just to clarify, crowds like GFC offer you 'leverage' - i.e. for $10 of your own money, you might be able to place a buy or sell order of up to $2000. This has an upside and a downside - the upside is that those microscopic swings in the market that are happening constantly, might therefore be exploited to actually make a measurable amount of money. The major downside of leverage is, if you get stung, you get really stung. Luckily with GFC (and I'm not trying to advertise them), it appears that you are never risking more than the money you put in - so you can't end up owing what you didn't decide to risk in the first place. You can of course, lose everything that you do put in - never gamble with what you cannot afford to lose!

One other thing you need to watch out for, is the spread. The price you can buy at is 3 points higher than the price you can sell at. What this means is, if you place a buy order, the price needs to increase by 3 points before you can break even by selling at the same price you bought at. This is the spread, as as I've found out, it's a killer.

So I've come to look at this like a sort of quasi-maths/stats problem: if you work from a presumption that the rate/price of the currency pair you are purchasing is *totally* unpredictable, is there a way that you could intelligently place buy and sell orders, in such a way that you could be in-profit, on average?

The orders you place are effectively 'positions' that you are taking, which remain open until you decide to close them. When you close them, then you either recoup the profit if the change has been to your benefit, or you are due for the loss! There are two major tools when you have placed a buy (expecting a rise) or a sell (expecting a fall) order. They are the 'limit' (a profitable price at which your order will auto-close), and a stop-loss (a loss-making price at which your order will auto-close).

So if goats are $10, and you place a 'buy' order on 10 goats with a stoploss of $7 and a limit of $12, then if the price rises to $12/goat, then your limit orders execute, closing the position and you get 10 x $2 as profit. If however the price falls to $7/goat, then your stop loss kicks in, your position is closed, and you lose 10 x $3. Easy right?

Well, I've been trying for a little while on the GFC practice account, and I'll put my hands up - I've given up.

I tried a couple of different strategies...

Chasing Momentum
This strategy was to "chase the momentum" on a tiny scale - place buy orders as the price was rising, and try to sell to cover those orders when the price had risen far enough to be profitable (and the reverse on the selling side - sell during little collapses, and then try to buy back and cover at a lower price). Great in theory, the problem being that sometimes the price reverses and never returns. So, it's going up, you stick on 2 or 3 buy orders, then it reverses before your positions are profitable, and keeps going down. You're then left with a choice - do you liquidate the unprofitable 'buy' positions for a small loss - or keep them open as the price plunges, and hope that it comes back up again?

Adjacent Hedging
This was an attempt to basically simultaneously place a buy and a sell order, both with limits and stop losses, and therefore make a tiny profit slice no matter which way the market moved. So if for example, you bought at 1.3000, then you were simultaneously selling at 1.2997. The issue here was that a limit order needed a minimum of 8 points distance, so on the 1.3000 'buy', the lowest limit you could have would be to sell and cover at 1.3008. As a result, the stop loss you needed on the 'sell' order had to be 1.3007, but as that equates to a 10 point loss, you need to make at least 11 points on the 'buy' order to be profitable. Which means the buy at 1.3000 needs a stop at 1.3011. And the reverse on the sell side. Ultimately, what it means is that the buying price going up, is going to bump the stop-loss on the 'sell' at 1.3008, but won't hit the limit to take profit on the 'buy' until the selling price gets to 1.3011 which is 6 points later. Or to put it simply, your stop-losses on both orders are several points closer to your current price than your limit orders, so the statistical likelyhood is that all the micro-volatility in the price will mean the price just drunkenly wandering about, nudging your stop-losses and accumulating you losses, and only rarely getting a limit that will make you tiny amounts. A good way to lose money.

My final attempt was based on something I noticed about the way the price behaved during a day of trading. The exchange prices wobble around set levels for a good while, not doing too much, and then at certain points, they 'break out' of the range they have been wobbling in, and take a quite visible 'step' upwards or downwards, where they then recommence their small volatile wobbles at a different level. So I came up with a new idea:

Capturing Breakouts
The idea here was to wait for the pattern of wobbling to emerge over time - look for a price that had been within a fairly obvious range for half an hour, or longer. Then, set a significantly large buy order several points above this range, and a similar sized sell order several points below - both with limit orders a few points away. The idea was that upon the market price undergoing a 'breakout' upwards or downwards, these automatically placed orders would 'capture' the breakout (while I cancelled the order in the opposite direction) and then as the breakout in price proceeded, the limit would kick in, and I would automatically collect the money. The big orders were a way to limit the need for trading - so that I would only have to capture 2 or 3 breakouts in a day to make the money I wanted. Safe concept, right? No. What I actually learned, is that there is no way to accurately estimate the 'wobble range' that a price will exhibit, so there is no way to safely set the order levels for the buy and sell. Similarly, there is no way to know how big a breakout will be, which means that often what happens is you get a little bust out of the normal range, which triggers the buy or sell order (depending on the direction of the bust out) - and then the price settles back into the range before it reaches your limit, leaving you with a very big bet that may be about to go badly south, taking a stack of your money with it.

In short, I've been unable to come up with a way of securing your positions, without the risk of leaving positions behind as the market price moves. The only guard against leaving positions behind is to always use stop losses, but my experience has been that stop-losses combined with the 3-point spread, effectively mean that you will, on average, lose.

I'd love to know if there is a way to exploit the random movements, but I can't help but feel that the spread between the buy and sell prices is the thing that tilts the odds fatally away from you.