May 3, 2008

Betfair Cash Machine – End of Test Review

I shall divide this review into two parts: the first, primarily factual, will record my actual findings.  The second part is subjective and deals with my opinions and impressions of the system and the eBook.

Findings: In the end I stayed with my original intention and ran the test over 28 days.  In that time I traded on 16 days, though two of these “days” were for ease and simplicity in posting and covered around 72 hours each.  I played 22 sessions: as these games can be played at any time this was, again, for ease of reporting and this is for information only.  The number of games I played was 231 and the system made a loss of £10.57.

In order to be fair to the author, this loss of £10.57 needs to be put into context.  Shortly after I started the test the bank was down by almost 30% and in order to protect it I reduced the stakes from £5 to £3.  Had I not done so the system would have made a profit of around £2.  This means that this particular “cash machine” would have churned out 12.5p for every day I traded, or 0.865p per game – I glad my high street cash machines dispense money a little bit quicker than this.  That aside, given the results up to that point, I think I would have been very foolish if I had not acted to protect the bank.  Also, if this had not been a test I would have thrown in the towel – or reverted to paper trading – when the bank was down about 30% in just two or three days.  Based on that, I feel that the loss of approximately £10, or just over 10% of the bank, fairly represents the profitability of the system and I shall recommend to Graham that this is filed under “Failed Systems”.

Moving on to my thoughts and impressions of the system and eBook: I don’t think it’s sufficient to simply state that I think this is a very poor product, I feel I need to give some reasons why I think it is so poor and also to give some examples. 

The eBook is quite short and contains the usual explanation of how to use Betfair (I wonder if the author would be doing buyers a favour if he left out this information.  They could then save themselves a lot of time spent in front of a PC, with their eyelids propped open with matches, earning 12.5p per day).  That the eBook is short doesnt particularly concern me – one of the most successful systems I use can literally be summarised in one line.  What does concern me is that the publication seems to me to have been dashed off in a hurry: terms like “slapdash” and “cavalier” spring readily to mind.

It may be that I have somehow misread, or misunderstood large tracts of the eBook and that there is, therefore, another explanation for what I about to mention (if there is I’d love to know).  Assuming the fault is not with me, my take on the book is that it has far more than it’s fair share of inconsistencies, inaccuracies, etc.  Possibly the worst error is an assumption made by the author that appears to me to be totally false.  If I am right this is extremely important, as the entire trading strategy is based on this assumption.  Before I deal with that point, there are two others that deserve mention.  First, the author states that “The odds are in our favour … … due to our extensive research through the Hi-Lo statistics…”  Ignoring the fact that the odds quite clearly were not in my favour when running the test, that was the sole mention of “extensive research”.  Apart from the author providing a link showing that it is, indeed, possible to download past results, I could find no evidence at all to lend support to the contention that any research had been carried out.  The clear inference I did draw, however, was that if any research had been carried out, it was in the form of back-testing only.  I was left asking myself whether the “extensive research” – months and months of it, maybe – was conducted in about 30 minutes.

My second point relates to the authors advice for when to close a position.  Common [market] wisdom is that the difference between successful and unsuccessful traders is all about knowing when to take a loss and close an unprofitable position.  The author is very clear when to open a position, but otherwise I found his advice vague and unhelpful.  The advice is on the lines that once the trade is opened and the odds move the “right” way, the position can be closed for a risk free trade.  He then says you may do better to let the position run for a bigger profit – he seems to miss the point that the profit is just as likely to be smaller and that an outright loss also becomes more likely.  Similarly he advocates that when the odds move the other way the bet should be left open, and then goes on to say that it could be closed, but generally he lets it run.  I would think that anyone who has not traded before would be very confused by this “advice”.  Any trader in financial markets who trades in such a fashion deserves to make a loss – unless a trader is very experienced indeed, there needs to be in place very clear strategies for when to take a profit, when to cut a loss and so forth.  These strategies I found sadly lacking.  Also, while the author makes it clear enough when to close a position for a “risk free trade” and when to “green up” he doesn’t make any mention of whether a loss should be for a “red screen” (roughly equal loss no matter what the outcome), or whether to risk a larger loss in the hope of breaking even.  Lastly on this point, he comments on the “wisdom” (my word) of greening up under certain circumstances and the likelihood of a particular outcome.  I don’t know which Betfair Games Hi-Lo he may have been trading on, but it doesnt seem to be the one that I was using: if I had ignored his advice and not “greened up” that £2 profit that I could have earned might have been even bigger!!  I suppose with hindsight I could have tried asking the author for advice, but when I see what appears to be a blind man directing traffic I tend not to show him a map and ask for directions.

Coming now to what I found most worrying: the author states that the system is based on an anomaly with the price and liquidity relating to this market.  While it cannot be ruled out that there may be an anomaly, my results call this into question and I can think of another explanation that makes much more sense to me: this explanation also fit’s in with how the prices actually move.  There isnt room to go into detail here, but I firmly believe that what the author claims is an anomaly is actually a function of Betfair ensuring that price and liquidity mirror as accurately as possible the actual probability of an outcome.  Given Betfairs resources, I assume that everything is rigorously tested before going “live”.  If an “anomaly” appears in every game and a similar “anomaly” appears under specific conditions at certain points during each game then I cannot believe that Betfair missed this or that it is an anomaly.  This is particularly difficult to believe when the “anomaly” is so obvious that, when pointed out, an eight year old will have no problem spotting it.  Am I to believe that Betfair, with all it’s resources, cannot spot and rectify something that obvious?  The evidence gleaned from this test suggests that if anyones reasoning is wrong it is not mine and, if I am correct, the whole system is based on a false premise.  Need I say more?

I think I have already rambled on long enough, I have (I hope) more than made my points and there isnt room to cover others I took issue with – believe me, there are others.  I will, however, make two final points before closing this missive.  First, I no expert but I have done some Internet based research and Ive learned from the sites/forums dealing with probability that all Betfair Exchange Games are “zero sum games”: so, unless there is a genuine anomaly, or in depth and extensive statistical analysis reveals specific situations that can be profited from (either of which are likely to require a lot of work to uncover) there will be losing streaks and winning streaks, but play enough games and the outcome will always be within a very small percentage of breaking even.  Unfortunately, this is before Betfair deduct their commission and doesn’t take into account that most bets will be matched at a price that is slightly worse than actual probability – if the probability is 50% then back bets will most likely match at 1.99 and lays at 2.01 – over time this and the commission will erode the bank until there is nothing left.

My last point is that, while the author mentions the “anomaly” that is used to signify where the opening trade should be placed, he makes no mention that, using recommended stakes (£5.00), this also means that only a very limited number of people can trade using this system before the odds will move, with the result that anyone who is not quick will have to trade at a less “commercial” price and so have any chance of a profit further eroded.  On the other hand, surely the fact that only a few people can trade using this system must ultimately be good news for those who cant get their bets matched in time?

You can get Betfair Cash Machine here:


Filed under Virtual Games/Casino by Michael


Previous post:

Next post: