It Doesn't Get Any Easier

If you have read my How to Read a Sports Trading Website article then you will have seen through those vendor articles telling you how easy in-play trading is. Trading tennis in-play is only for syndicates with court-siders plugged into algo-bots. Sitting at home with a delayed video feed will do little more than earn trading profits for the syndicates and subscription profits for the vendors.

My last article, The Paradox of Skill, received a reply from a commenter, confirming the paradox of skill from their own experiences. To recapitulate, the paradox states that as the skill level of a group of traders in a market increases, the abilities of the traders will even out such that all traders will have an equal chance of winning - i.e. luck plays a greater role.

Imagine an idealised market with only two traders, one is an expert and never makes a mistake. Whenever they back a selection the price falls and when they lay a selection the price rises. The other trader in the market is a beginner, they are always on the other side of the expert's trades, the wrong side. In our idealised market, the expert always wins and the beginner always loses, therefore the expert overcomes commission costs and makes a profit. 

The beginner goes away to lick their wounds and decides to learn how to build better trading strategies. They are successful and their skill level is now equal to that of the expert trader. Because the two traders are equally skilled and there are no other traders in the market, the two traders end up pushing money backwards and forwards, between themselves, whilst paying commission.

Sometimes the former beginner gets slightly ahead only to mean-revert and sometimes the expert trader performs in a similar manner. That is expected variation and in the long-run their net profit is zero minus the commission paid to the exchange.

A third trader - another beginner - joins the market and for a brief time the two experts compete for the losses incurred by the beginner. Eventually, that beginner learns how to be an expert trader and once again all traders are now losing to the tune of the commission rate.

Other traders may join, some become expert, some leave because they don't want to learn how to become experts and maybe a few are wealthy enough not to care that they lose money. It is the careless that the experts depend upon for their profits but if every trader is an expert then only luck can create a winner.

As the number of beginners in the market increases, the probability of a beginner getting lucky increases. A lucky trader might post a short-term profit (not to mention, starting a boast blog) only to mean-revert in the long-term and lose.

We can use the preceeding thought experiment to explain the consequences of trading courses. Imagine an educator has cornered the market for trading courses and has personally trained each and every trader in our hypothetical market. All traders have the same expert knowledge and so they too will push money between each other and lose the rate of commission in the long-term. We might call that the paradox of education, whereby being as equally educated as everyone else garners you with no advatange in life.

Why then would a vendor train people up to trade as they do? It can only be for one of two reasons. That the educator is not training his students to trade as he does but is misinforming them solely to get them to enter the market so that he can take a contrary position in the market. Alternatively, the educator doesn't care what these traders do because his main income is from teaching and selling other products.

You might ask if this paradox occurs in financial markets too. Indeed, it does. How do financial traders get around this problem? Well, they use high-frequency trading, dark pools and keyboard readers to intercept discretionary trader orders. All so that they may front-run other traders' orders before they go on to the exchange. Volume is bought ahead of the discretionary order by an algo-bot that has been tipped off to the trade and acts as a middle-man, taking a cut.

If all else fails, traders cheat. LIBOR and forex fixing, and the sub-prime mortgage scandal whereby the mortgages of people who had no chance of repaying the loans on their houses were sliced and diced with triple A loans to create CDOs that could be palmed off on the unwary (i.e. Deutsche Bank). 

Today, financial markets are so informationally efficient and so much processing power is thrown at the trading problem that in order to get the bonuses that traders crave for there is no option but to use underhand methods. Hence, we now have court/track/pitch-siders and match fixing in sports.

The commenter wrote:

I wanted to say thanks for writing your latest book, and this seemed as good a place as any.

I've been trading algorithmically for 4 years and am always happy to buy any book that may be able to give me a sniff of an edge. I can now confirm that you are telling the truth when you say this book doesn't offer one, but it's reassuring to see so many similarities between approaches, and I would recommend it to anyone taking their first steps into algorithmic trading.

In terms of this post, I can definitely confirm things are getting tougher all the time. My simple rules based code from 4 years ago has got much more complicated over time, but I'm currently making less with machine learning models than I was with the simple rules a few years ago. I always thought this an inevitable reality, but a few things do worry me. First, the introduction of the fee to use the betfair API can only reduce the number of new traders. Second, Betfair have a virtual monopoly in the sports exchange market, and whilst I have nothing against them, we are all at risk of changes to their terms (I may be doing BETDAQ a disservice, i haven't used their API for a few years). Whilst 10 years ago it may have seemed that the sports exchange was the future of betting, now it's not entirely clear if the sports exchange has a long term future, and it worries me every time I logon to betfair and realize that to a new user they're primary a sports book these days. This is one of the main reasons this remains a hobby for me.

It appears that I've stumbled upon a community of like minded people, so whilst I have no desire to share what I think are my edges, if anyone's interested in sharing some notes, let me know!

I am always happy to hear from other algorithmic traders even though edge is never shared. Algo-trading is not the glamorous (glamour seeking?) side of trading. Manual traders can play at being Bud Fox in Wall Street or Jordan Belfort (Heaven forbid) in The Wolf of Wall Street. Algo-traders have no pin-ups. With no ego to massage, algo-traders tend not to boast about their trades. "Getting it quietly" is definitely the catchphrase amongst algo-traders.

To be honest, I look forward to the day when all markets, both financial and sports, are filled with expert (most probably AI based algo-bots) traders who neither win nor lose. D-Wave's quantum computers may be edging us towards that day. Financial markets will return to their originally intended purpose, providing capital for economic growth and the eradication of poverty. Sports exchanges will be places where the poor don't lose their last pennies.

Does the sound of that bore you? Not I. The AI of the future, which creates efficient markets that are difficult to lose money in will probably be powerful enough to ensure that nobody needs for anything and there will be better things to do than risking money for gain. Maybe money itself will become obsolete. Until then, you have to hope that you are on the side of the informed and not the uniformed (or misinformed).

Of course, some sports exist solely as a conduit for gambling. Horse racing being the most prominent. There is nothing I can say that would console those who would mourn the loss of horse racing, but if given the choice I'd rather live in a world without want of food and shelter than one with a want for horse racing.

No comments:

Post a Comment