fred77's Blog

Sorry, I don't know a decent one. But I use and am happy with my virgin cable.

If you are actively trading you want your own dedicated connection.

I wouldn't even consider testing a bot if it was behind a home connection shared with "6 computers, a PSP, two smartphones and a Nintendo Wii"

Teenagers are the worst, they suck up every last bit of bandwidth with their file sharing, utube and so on.

lol no, I don't even have a framework.

IMO there really is no need for a library of code. Code just demonstrating how to connect, login, etc for the common languages would be adequate. Once you know how to connect to the API with your chosen language then you just do it, variables are passed and returned just like calling any other function. If you want you can write little routines to interact with/wrap/abstract the API in the way that you choose. The API services mostly behave as the manual states.

I guess at the outset, the sheer magnitude of API functions and available data to juggle can be overwhelming, and wanting to see how other people choose to handle it is only natural :)

Sooooo, basically, you'd like an off the shelf bot framework for free? mmmmm

Yes it is their way of throttling. Here's a relevant bit from an old email (might have changed by now):

The thresholds are applied in a historic 60 second sliding window and the blacklisted in period is also 60 seconds - during this window no API calls will be processed [with exception of Cancel] and an blacklisting return code is sent back to the calling application. You can cancel during a blacklisting window, but this does contribute further to the length of the blacklisting window. You can call the ListBlacklistInformation call to get details of blacklisted API call and how long more the user is blacklisted for.

The below values are default values and we do have the ability to change these on a per user basis, but in reality.

The threshold [per minute are] are

PlaceOrdersNoReceipt = 1000 [count of all calls made to the PlaceOrdersNoReceipt ]
PlaceOrdersWithReceipt = 1000 [count of all calls made to the PlaceOrdersWithReceipt ]
Combined = 1500 [count of all calls made to the api]
Any=1000 [count of calls to any api call other than PlaceOrdersNoReceipt & PlaceOrdersWithReceipt]
ListBootstrapOrders=50
GetEventSubTreeNoSelections=50

API fully down at moment. Guess it's all still to do with Monday's upgrade.

There's quite a nice VB2008 tutorial / work through appearing in the BDP forum. Everything is free. If I was running windows I'd be very tempted to go for that.

oh well, doesn't seem to have cost me too much (not that I can really tell with my current volatility!). Seems that a few of the api servers went a bit wonky for a while, giving stale data and other problems. My bot wasn't recording the issues with stale prices, but I managed to clock up 57 BET_IN_PROGRESS error messages while placing bets. My bots did initially stop and report the unusual error, but I decided to override it thinking that it might be a money spinner if other bots were having issues. With hindsight I should have looked much closer at my log files before making that decision...

email notification definitely isn't working anymore

mmm I'm depressed, can't stop drinking, doesn't matter how much money I make, I'm still fed-up

The betfair API is shafted too!

Probably is about time we had another little poll to see what our favourite bot writing languages are. No change here, I'm now into my 3rd year botting with PHP.

LOL Fred. Us "old" boys don't like change, especially when learning a new language means a LOT of extra work which at the end of it all does exactly the same as what we already have. Learning something new is even less appealing when you only fart about with programming as a hobby/early retirement dream.

Dunno if you read my post above (not sure which one), but i THINK you can now use PHP to write stand alone Java apps. So the theory is that you can write Java code using the PHP syntax, or something like that. Google "PHP JVM", i'm sure there's something of interest.

-birchy

Yes I don't like change - I just occasionally think the grass is greener elsewhere. I'd really hate to spend ages learning something, developing my code, and then when it comes to the crunch find out it's no better (or worse) than what I currently use. I had a quick google for PHP JVM, it's not really of interest to me as I'm not interested in java bloat or IDEs.

I wouldn't mind a php compiler, last time I looked they (roadsend compiler) didn't fully support PHP's full library of functions though. I've just found that phpcompiler.org is finally out as a beta, it includes some optimisation, I'm going to see if that is of any use later, as shaving off 0.01 milliseconds here and there will make me feel happier even if my bots don't perform better lol

It's probably for the best that I've never ditched my beloved PHP! I too prefer rapid edit, test, edit, test. Though that could be down to the fact I introduce so many errors that they become hard to spot if I leave it too long between test runs lol

http://www.highstakes.co.uk/shop/product.php/12204/0/

Yes you'll probably regret it, but curiosity has to be satisfied :)

I think I paid £30 for it. As I said, not much use to any experienced bot writers but probably of interest to beginners. Basically, if you have succeeded in satisfactorily gathering prices, applying strategy, and placing bets, you won't need the book - but that said, who knows, the text may inspire your next fabulous moneyspinner...

I don't think there was too much more in the way of strategy, without looking at it again I couldn't say for sure. The book was well padded with the PERL examples.

Brendan, the book recommends the use of linux for stability, and the example bot is written in perl, and uses mysql. It demonstrates using information from the racing post postdata table to roughly rate and price up horses and then bet/dutch at the business end of the market.

nashenden, as you've been a programmer before I reckon you could easily pick up enough to get by. The learning curve quickly levels off once you get working connection code.

An awkward bit is that you want to sell your software, which necessitates it being of merchandisable quality - imagine all those different platforms to support and lots of moaning from unsatisfied customers.

Seems slightly bizarre that you would want to sell your golden egg laying goose...

I got it a fortnight ago or so, when I saw it mentioned on bf forum.

It provides a gentle walk through building a simple bot. I read it cover to cover, and it was ok, though I certainly wouldn't recommend it to any experienced bot writer, as you won't learn anything new.

I was reading away thinking it's going to get to the exciting bit soon, it never did though! Upon reading certain bits I found myself thinking why on earth would I want to do it like that?

It was refreshing in the fact that all software used was open source. MS Excel didn't get a mention in the whole book :)

Also, I should point out, I don't claim to understand the premium charge. I don't think it will affect me, but seeing my accounts rise an awful lot this past week has naturally made me somewhat concerned.

Regarding your idea jpl, previously I've already done similar with a few sports as part of my drifter/steamer type experiments. Breaking even is indeed very possible using such simple methods.

I'm still more interested in making a strategy that will arb £X in a chosen direction upon command so that I can shift unexpected profits without any further worry.

It may well be that my current bots are so volatile that they're generating enough commission to pay for any super positive trends (and they are just as likely to take a massive dive). I just don't want to find myself paying for good spells and not getting any compensation for the bad spells.

If you do find a way to "reliably and automatically move funds from betfair to betdaq" you would surely be better off running the betdaq half of the process only (even though it means you pay more PC at Betfair).

-jpl

It would avoid premium charges because there wouldn't be an overall profit at betfair. The idea is that the betfair account would have lost money on arb bets while the betdaq account gained it. It's a bit awkward though, because like all the best betting strategies it's much easier to do when you know the outcome of an event beforehand lol

I don't specifically want to share funds, I just want to have an automated arbitrage tool available that I can run in order to move excess funds from betfair to betdaq, to guarantee that I won't pay any premium charge after having a good winning run. Paying a bit extra commission to do the trade isn't an issue if it's gonna save a bigger chunk and/or give me peace of mind.

As per this thread title, after having a very good week last week I found myself wanting a work around last night (Sunday). I've no idea if I've actually incurred charges but I would have gladly taken steps to ensure I hadn't. I have been looking at arbing but still can't figure out how to reliably and automatically move funds from betfair to betdaq.

I find that any bot that is capable of good winning streaks is usually capable of much better losing streaks ;)

Looking at that sample I gave, I imagine something like this would work:

// initialise an array to hold bets
$betarray=array();

//add a bet to array
$betarray[]=array('Order'=>array('SelectionId'=>$id, 'Stake'=>2, 'Price'=>$price, 'Polarity'=>$polarity, 'ExpectedSelectionResetCount'=>$esrcount, 'ExpectedWithdrawalSequenceNumber'=>$ewsnumber, 'CancelOnInRunning'=>true, 'CancelIfSelectionReset'=>true, 'WithdrawlRepriceOption'=>2, 'RestrictOrderToBroker'=>false, 'PunterReferenceNumber'=>0);

//add another bet to array
$betarray[]=array('Order'=>array('SelectionId'=>$idb, 'Stake'=>2, 'Price'=>$priceb, 'Polarity'=>$polarityb, 'ExpectedSelectionResetCount'=>$esrcountb, 'ExpectedWithdrawalSequenceNumber'=>$ewsnumber, 'CancelOnInRunning'=>true, 'CancelIfSelectionReset'=>true, 'WithdrawlRepriceOption'=>2, 'RestrictOrderToBroker'=>false, 'PunterReferenceNumber'=>0);

$result = $BDSecureSoapClient->PlaceOrdersNoReceipt(array('placeOrdersNoReceiptRequest'=>array('Orders'=>array($betarray), 'WantAllOrNothingBehaviour'=>false)));

mmm I'm confused now, it's late, maybe this instead

$result = $BDSecureSoapClient->PlaceOrdersNoReceipt(array('placeOrdersNoReceiptRequest'=>array('Orders'=>$betarray, 'WantAllOrNothingBehaviour'=>false)));

I haven't had any problem with betdaq API usage, I thought their limits were quite high (according to an old email I have from them - around 1000 calls per minute for most services apart from ListBootstrapOrders and GetEventSubTreeNoSelections which are 50)

Lovely and quiet on here, yesterday was the big day, surely one of us must have a woeful tale of premium charge deductions?

Well, I would never have thought of that. I didn't realise .com and .co.uk were different. I guess I must have signed up with .com as that comes up first in google lol. I didn't twig that that was why you were pointing your script at .co.uk, now I see. Enjoy your walk :)

I think betdaq's api is slow compared to betfair's. Not terribly slow, it may just be me being a speed freak! I'm not quite sure where betdaq's servers are (well I don't know where betfair's are to be honest!) but I find getting prices takes around 10-60ms more than betfair and placing bets (without even waiting for matching) takes around 200ms (but upto 1600ms) longer.

Yes, I had a go at scraping betfair and found it around 10ms faster than API. I don't scrape anymore though because of data charges.

I'm still not sure why my code is failing for you. It should work as is. Perhaps removing any cached wdsl files might help.

At least you figured out the problem. Simple problems can be the hardest to spot.

I'm starting to look at betdaq now, trying to figure out how to arb, I thought it would be fairly easy but getting both sides matched is rather hard, and is getting harder with betdaq api running like treacle.

Hi ossi. I can confirm that the given code still works ok for me. The bit where I add the soap header is a bit odd, but I couldn't get it to work any other way (adding the header along with attributes). Are you using a recent version of PHP?