Wednesday, 27 September 2017

Botters and Spotters

In my always humble opinion, the Q and A series from @Statsbet reached a peak a couple of weeks ago, and further evidence that it is now on the decline came yesterday with the disappointing news that the latest in the series was to be with an unknown - someone called Betfair Pro Trader. 

Talk about desperation and scraping the barrel, and rarely has the expression 'jumping the shark' been more apt.

Long time readers will know that I have never heard of James Butler of Betfair Pro Trader before, nor have I ever bought any of his books called Betfair Trading Techniques: Trading Models, Machine Learning, Money Management, Monte Carlo Methods & Algorithmic Trading.

But I'm always ready to learn something, in fact I have a saying that a day is lost if I haven't learned something new, so I took a look at the thoughts of this previously unknown trader who specialises in automated betting.

Number 2 blog! What a great question. I wonder why I wasn't asked the same one? 

The question of staking plans was asked, with James' reply below:
James could be accused of selecting Kelly because of the Irish connection, but it's a mathematical fact that the Kelly Criterion is the optimal staking method for maximising profits. However, in the real world of sports betting, Kelly has a couple of shortcomings, specifically that the precise edge is seldom known and the size of stakes and thus draw-downs using full Kelly means that most people will abandon it sooner rather than later. My 2008 post on the topic holds true, although the final paragraph's conclusion that:
 Increasing this [2% stake] to 3%, or occasionally 4% on an especially good play, is reasonable
was written in the days when many markets were still free of courtsiders, (or 'spotters' as James calls them), and opportunities were more plentiful for the skilled trader.

About level stakes, I would say that this is the only way to validate whether or not a system is profitable, and for comparing one strategy against another, but for building a bank, it's far too passive. My approach is to increase stakes as the bank builds. 

R J Miller uses a similar method to me, but with a plateau system to determine when he increases his stakes:
I bet 2% of my bank and continue to flat bet until my bank grows by at least 25%. Then I recalculate the 2%. Thus, if I started with $10,000, I would bet $200 a game until my bank grows to at least $12,500. At that point, I would refigure my unit to $250. I bet flat until I reach at least $15,625. My actual risk-reward ratio doesn't get too high. The other thing that I do that is unique and rather arguable is that I never lower my bet. Remember, if you vary the bet, your breakeven goes up. At a 2% unit and three decades of experience, I am comfortable that I can ride through a losing streak. If you lose 10 games at $200 a bet and lower the bet to $180, you must win 12.2 bets to get back to even. This is how I get my 100% annual ROI. My actual bet goes up during the year as my bankroll reaches higher plateaus.
I really don't see the point of the plateau system here. We all have spreadsheets these days, and 2% of $10,000 is as easy to calculate as 2% of $10,123, so why slow down the bank growth unnecessarily? Each time the bank hits a high, increase your stakes.

It's also neither unique nor arguable to never lower the stake once it's reset.

If you're curious about the example on the left, and why the stake size is static despite wins, it's because many games are played on the same day, not necessarily simultaneously, and I (hopefully) raise the stakes at the end of the day for consistency. Some days I might be home and in a position to raise the stake, but more often than not, I am having a life. I'm generally past the stage of checking results in real-time.

I jest about Mr. Butler of course. His expressed opinions about the realities of betting match my own, and he's always worth a read. You might not like what you read, and indeed James writes:
My wish is either to point people in the right direction or to frighten them off. I regard putting someone off betting/trading as of equal merit to teaching people useful skills.
How many fortune seekers are actually dissuaded from pursuing betting is debatable, but close to zero I would say. The nature of a gambler is to pay far more heed to the "winning is easy" garbage churned out from easily discredited sources if you really want to check, than to more rational thoughts and the logical conclusion that:
It's not impossible to make a profit, but it's hard work, and requires a certain disciplined mindset to achieve.

A jump in the number of hits to this blog over the past couple of days, and with Tweets like this one below, today could be an "enormous" one too.


Dr Tsouts said...

Speaking of banks and staking what do you suggest to those who follow many tipsters? Keeping a common bank or an individual for every tipster?I guess the second is better but we all have a life eventually...

James said...

If Master Cassini had actually read Betfair Trading Techniques he would have access to a spreadsheet and simulator to test level stakes against other money management systems.

Even a fractional Kelly kicks sand in the face of level stakes.

Whilst useful for testing a strategy, level stakes is woeful for building wealth.