Remove this ad
avatar

birchy

Betfair Elite

Posts: 593 Member Since: May 11, 2008

Lead

December 16, 2012 17:08:37

Tags : :

Since finding a couple of profitable strategies over a year ago, I've done very little towards developing anything new. Although I'm a strong believer in "if it ain't broke, don't fix it", I can't help wondering if I should be developing other strategies "just in case" the current ones take a turn for the worse.

What with another career promotion and the 8 years it took me to reach my "Eureka!" moment, this last year or so has been a welcome rest break and I consider the ~£3K a month income as back-pay for all those years of trying and failing. It's true that I have become complacent, however I've also gained 20+ hours per week of my spare time. Although my strategies are solid and the only way they would start losing is if Betfair make major changes to the betting exchange model, it's not an impossibility.

So the question is...should I be devoting some time to developing new strategies or should I continue enjoying the welcome break?

www.bespokebots.com

"This time next year Rodney, we'll be millionaires!"

Quote    Reply   
Remove this ad
Remove this ad
avatar

pic

bot addict

Posts: 63 Member Since:July 28, 2011

#1 [url]

December 17, 2012 09:16:59


hi Birchy,

first of all kudos for your great achievement!
Having been working for almost 3 years on betfair bots (a *lot* between 2009 and 2011, a lot less the last year and half for lack of time), I know it is not easy at all.
I haven't yet found any "Eureka!" like you but I still have a bit of hope ...

To reply to your question: I would say: yes.
If I were in your comfortable situation, after a long rest as you did, I would start to work, very very easily, on something new, "just in case". Perhaps, you could also consider something independent from betfair, therefore you build a lifeboat in the unlucky event that an unexpected event causes troubles to your strategies or with betfair in general.

In the end it all depends on your nature, I suppose.

Quote    Reply   
avatar

birchy

Betfair Elite

Posts: 593 Member Since:May 11, 2008

#2 [url]

January 12, 2013 16:31:00

TBH, I know it's probably a case of "the more, the merrier!" but winning strategies are not easy to come by and I know from experience that a few hours per week can quickly turn into an obsession.

www.bespokebots.com

"This time next year Rodney, we'll be millionaires!"

Quote    Reply   
avatar

blackmagic

bot addict

Posts: 70 Member Since:June 26, 2008

#3 [url]

January 12, 2013 17:35:11

I don't think its complacency to not be working on them. I suppose if anything with strategies its really about knowing exactly what you're edge is and working out if there are other markets that you could apply it to.
I was reading an interesting trading book a while ago and one of the points made was when you have a working system then dont look for it to start trading more frequently, look for it to trade bigger size on each trade you do. Mapping that to betting i would say volume is the key to pnl but maybe either getting the strategy working to betdaq or else seeing if you can get any luck on the dogs or maybe some other market.

Quote    Reply   
avatar

pic

bot addict

Posts: 63 Member Since:July 28, 2011

#4 [url]

January 13, 2013 10:44:57

I see you point birchy but I also think that in your position, with something that really works, you should be well protected against obsession :)

Quote    Reply   
avatar

birchy

Betfair Elite

Posts: 593 Member Since:May 11, 2008

#5 [url]

January 13, 2013 22:22:12

The strategies I use are already working on *all* sports, so I can't really increase volume in that way. Liquidity is my main limiting factor as increasing the stakes does not give a proportionately bigger return. In fact, minimum stakes give the best ROI and the ROI reduces as the stakes are increased, resulting in similar monetary profits but with higher risks. That is actually a useful feature as it allows me to keep my commission rate just under the premium charge threshold.

Regarding "spare" time, I have a number of engineering projects on the go, including converting a milling machine to CNC, so am basically utilising the time I *would* have spent programming on other things. But I must confess that I often think I should give up my day job and then I'd have plenty of spare time. But there's always the doubt that if the betfair bots go wrong, then I'd have no income and may not be able to get a job like the one I have now...

www.bespokebots.com

"This time next year Rodney, we'll be millionaires!"

Quote    Reply   
avatar

pic

bot addict

Posts: 63 Member Since:July 28, 2011

#6 [url]

January 16, 2013 19:14:34

maybe working 80% or even 60%. You would keep the job and you would have much more free time. Is it an option for your company?

Quote    Reply   
avatar

jpl

bot addict

Posts: 82 Member Since:May 10, 2008

#7 [url]

January 16, 2013 23:20:07

Hi Birchy,

For comparison, I found a (very slightly) profitable strategy three years ago, and have been running it (almost entirely unchanged) across football and some other sports ever since then.  Like yours it works best on small stakes (a few pounds), and so can't be scaled easily.  It places tens of thousands of bets every month, but the average margin on each bet is a fraction of one percent.

My experience is that its profitability has gradually eroded, presumably as other people become better.  Net profit was £3,400 in 2010, £2,600 in 2011 and £2,000 in 2012.  Your mileage may vary of course, but I'm guessing you'll see something similar i.e. gradually eroding profits as markets become more efficient.

That's unless something more dramatic happens e.g. a previous strategy was sensitive to the introduction of cross matching, so saw its profitability fall away massively as Betfair rolled that out across more and more markets (a few years ago now).

Quote    Reply   
avatar

birchy

Betfair Elite

Posts: 593 Member Since:May 11, 2008

#8 [url]

January 17, 2013 23:05:43

Hi JPL, good to see you're still lurking!

Your strategy sounds very much like my old greyhound bot I was running about 4 or 5 years ago. It started off with about £2K profit in the first 6 - 8 months and then flat-lined for about 6 months, which was when I stopped it. But that was in the days when I was running test bots on my desktop PC, so they were never left on 24/7 and very often missed some races because I hadn't got the machine on.

Current strategies net £2.5K a month at £2 stakes and around £3K at various stakes up to £5 and that's been consistent month to month since December 2011. I've got my mind set on winning enough to buy a small house and then renting it out. If I achieve that, then the betfair strategies would not be the sole income and I could at least consider reduced working hours and possibly retiring before I turn 40. ha ha.

I really should look at implementing my main strategy on Betdaq as I see no obvious reasons why it would not work on ANY exchange.

www.bespokebots.com

"This time next year Rodney, we'll be millionaires!"

Quote    Reply   
avatar

worthapunt

rookie botter

Posts: 15 Member Since:January 29, 2012

#9 [url]

January 21, 2013 20:57:34

First of all well done birchy! The fact that you've achieved that does leave some hope for me and the rest of us. I was wondering how you arrived at this strategy? I'm obviously not asking for you to give away your edge but I was wondering about the process of developing a strategy. I ask because I tried a few simple things, which were unsuccessful, then dried up. I read that you spent time developing other peoples bots (for peanuts), did that help? Maybe you applied some financial trading strategy or was it just a flash of inspiration from all your experience? Anyway, I understand if you don't want to say too much about the issue.

With regards to your current circumstances my suggestion may be boring and sensible but, rather than retiring early, why don't you try and get a job you want to do? Of course that might mean training courses or something but you seem like a pretty motivated guy. Also, I think there are sports trading firms out there and surely you'd have something to offer them! Google 'siam football modelling' for an example.

Quote    Reply   
avatar

birchy

Betfair Elite

Posts: 593 Member Since:May 11, 2008

#10 [url]

January 22, 2013 00:14:25

Ironically, I took enough time away from betting and botting to forget everything I had been doing, so came back with a completely fresh mind. I actively avoided reading any betting and strategy type forums as they only cloud your judgement and turn you into another losing punter. It's also very easy to over-complicate strategies and end up wasting months (or years!) "tweaking" the plethora of settings you *think* you need. If you take an average bot and multiply the combinations of possible settings, it will soon become apparent that it could take about 9 million years to test them properly.

My strategies are very simple and often overlooked because they're too obvious and "everyone will be doing it". I was certainly guilty of having that mindset. I didn't fart about back-testing or compiling statistics, etc. And I don't do any trading because that's how you end up paying premium charges. I wrote a strategy and threw it into the live arena, which is, IMO, the ONLY way to truly test a bot. The biggest lesson I learnt is that volume is king. Barring programming bugs, I will no longer retire a bot until it has at least 5000 settled bets under its belt. More often than not, it takes 10,000+ settled bets to show it's worth. In essence, you create a line graph of your P&L and if there are spikes in the graph, you haven't tested it for long enough. It's basic stuff, but often overlooked when you're -£900 after 3 days and wondering WTF has gone wrong. Big balls are required as, even with a solid strategy, the swings can be large on a "gambling" bot.

Regarding the early retirement/working reduced hours/changing career/etc, I do actually enjoy my job in engineering. If I hated it, I would of jacked it 6 months ago. The main reason I'm starting to consider other options is because the company I work for has become one of those places where there are as many managers and directors as there are "workers". We've got all these little hitlers who are jostling for positions and think that a production factory will be more efficient if they put everything into a spreadsheet and make all the "cretins" work harder. It used to be a good atmosphere where everyone had a laugh and maybe took a few extra breaks but the work got done, and to a high standard. Morale is at an all-time low and production is suffering because of it. TBH, I wouldn't be surprised to see the company fold in the next year or two. So I might hang on for redundancy. ha ha.

www.bespokebots.com

"This time next year Rodney, we'll be millionaires!"

Quote    Reply   
Remove this ad
avatar

worthapunt

rookie botter

Posts: 15 Member Since:January 29, 2012

#11 [url]

January 22, 2013 17:23:28

That's interesting insight, maybe I'll look at resurrecting some of my ideas. I agree playing with parameters can take forever, personally I am wary of anything with more than one decimal place. I'm currently trying to model football results to identify value, slow return though...

The 'too many chiefs, not enough indians' syndrome seems to affect most companies at the moment. I did a placement at a factory that decided that the most important thing was to implement a new software stock management system which was in no way better than the bespoke system that they'd had written 20 years ago. It went bust a few years later. I blame MBAs and their ilk.

I only suggest alternatives to betfair bots because of the 'black swan' risks, like a UK government deciding to ban online gambling (happened in the US). In that case you'll want a back-up plan.

Quote    Reply   
avatar

toy

rookie botter

Posts: 15 Member Since:June 2, 2008

#12 [url]

January 23, 2013 20:19:17


I love Birchy's story, this is a great example of persistence and the will to succeed - they should teach it in schools! It got me thinking that botting seems to be fairly insular and wondering how much more profitable it would be to have a team working on a project.
I understand why people keep their strategies close to their chests and people are always looking for an edge...
Has anyone here collaborated, I'm thinking that the amount of combined years of experience that is on here in programming, market analysis, market risk and market value and strategy evaluation and testing - it would be a great idea to get a project going. Does anyone have any ideas on how something like this could be managed or indeed how much interest there would be?

Quote    Reply   
avatar

birchy

Betfair Elite

Posts: 593 Member Since:May 11, 2008

#13 [url]

January 28, 2013 21:24:50

I joined a collaboration project a few years ago with a few people from a poker botting forum but it died a death quite quickly because everyone seems to think that their little piece of knowledge or area of expertise is more valuable than everyone elses'...so they refuse to share anything useful (i.e. none-public) until they receive something special. It ended up being a multi-way conversation with nothing new being learnt. IMO, none of them really had much to offer and had only joined because they thought they could leech off everyone else. What you'll find is that anyone who is "successful" won't want to join a project for fear of being leeched...

www.bespokebots.com

"This time next year Rodney, we'll be millionaires!"

Quote    Reply   
avatar

worthapunt

rookie botter

Posts: 15 Member Since:January 29, 2012

#14 [url]

January 29, 2013 19:42:41

I think it would be pretty hard for collaboration to work. The main problem is that money, the internet and trust don't really go together. I have read about people forming companies to bot on betfair though, that's probably the best way, although they probably know one another so trust isn't such a problem.

Quote    Reply   
avatar

blackmagic

bot addict

Posts: 70 Member Since:June 26, 2008

#15 [url]

January 30, 2013 12:13:51

I would say that collaboration is difficult as you often find a working idea is simple and once it works its hard to add to it. I think this is what birchy is finding in regards to not changing his working strategies. 

I would love to know how you guys make money, i think i have an inkling given the things you have said about them being impossible to backtest and not being easily scalable. I spend most of my time doing statistical research and backtesting strategies. I have found this to be successful but only when i spent a lot of time reconciling real trading with sim. It is very easy for these to be subly different and i makes all the difference to your real pnl. My main break through was once i had found all the kinks to this, of which there were many, I then found the strategies i got from the simulator actually made money in reality. The upside to this is i can now tune them and change them without having to wait weeks/months to see if the idea is good or not. 

Quote    Reply   
avatar

toy

rookie botter

Posts: 15 Member Since:June 2, 2008

#16 [url]

January 30, 2013 14:53:25

You are all right, trust would be the major issue for any project being worked on my multiple people. Conceptually I like the idea of open source style projects where everyone is benefiting from input. Although Birchy must have a real sense of achievement, there must be a way to work together so that it doesn't take 8 years at 20+ hours a week (I hope you had a holiday) to start seeing real returns. When a team could do the same work in a fraction of the time and potentially realise larger profits.

What about a model where you have a team working on a project that is funded by investment, like a hedge fund where the investors get a return on their capital and the project team take a fee proportional to the profit (e.g. 10% of return). Plus team members could be financially invested too. Is there enough edge on the exchanges for this to be viable.? I'm just resurrecting previous thoughts and ideas I have had, and that's what they still are. But I'm sure if it can be done it will be done by someone if it isn't already.

Quote    Reply   
avatar

worthapunt

rookie botter

Posts: 15 Member Since:January 29, 2012

#18 [url]

January 31, 2013 22:22:56

With regards to the existence of edge, I imagine there is (an edge) somewhere. Look at the traditional financial markets, they've been around for decades and people still manage to find an edge. The main difference is the ludicrous liquidity of the stock exchange makes even an edge of 0.00001% worthwhile.

I think the difficulty in getting this sort of thing organised is the 'chicken and egg' order of it. If someone knows an edge, they're probably busy programming it. If you don't have an edge you don't have anything to program.

As I see it there are four main areas where people can contribute to such an endeavor:

1) Betting strategy, in particular experience on exchanges.
2) Programming, implementation of a strategy.
3) Data, for strategy identification and back testing.
4) Bankroll, the trial and error nature of developing strategies will probably require the (initial) loss of capital.

Really I only have reasonable skills in 2) and even then I don't know OOP.

Quote    Reply   
avatar

pic

bot addict

Posts: 63 Member Since:July 28, 2011

#19 [url]

February 1, 2013 01:54:52


@blackmagic: I have an approach very similar to yours.
I have a lot of testing data and I'm able to back test most (but not all) of the strategies that come to my mind. When I have an idea, I wrote the code and then I test it on the data I have, if the result is positive I may consider to try it live, otherwise I nix it. My testing platform is based on "healthy pessimistic principles", so it is unlikely that something which works well there could give worse results live, rather the contrary.

Some of the problems I see with this approach are:
1. you need to collect it a lot of data before you could adopt it,
2. collecting the data in not easy with the limits of the free API (you are going to "throttle" a lot) and also it should not interfere with bots which make bets. My data is not "good": I had a statistician looking at it once and he didn't find it so good but it is the best I can get so far.
3. also it is sometimes an issue to work with a lot of data. My tests take hours and moving Gigabytes of files between machines is also slow.

Now the funny thing is that I haven't made any money with this approach (until now?) but the last year I add some little satisfaction from a strategy that I hadn't back-tested because I didn't have the necessary data. More recently I started to lose money with this strategy, so I back-tested it and I've adjusted some parameters and I thought to be set. Unfortunately I'm still losing money but I'm sure it is not a problem of the testing platform which is accurate enough (I know that because I can always verify that real results and tested results coincide, and I make often this check, especially when there are troubles)

Quote    Reply   
avatar

theshadow

rookie botter

Posts: 9 Member Since:February 3, 2013

#20 [url]

February 4, 2013 18:56:14

I think that you should always look to evolve and try to stay ahead of the game. Its very easy to sit back and think that everything is going well until a big shift in the way the market operates ..... im pretty sure thats what has happened to HMV, Blockbuster Jessops etc

If i had a winning strategy i would just sit back and watch the profits grow, more in relief that i had finally found something and could stop searching.

My first aim to just design a bot, im not a programmer although i am interested in it so this is a good way to learn Python.

Its going to be a long road but its a hobby so its enjoyable.


Quote    Reply   
Remove this ad
Add Reply

Quick Reply

bbcode help