General Betting
There is currently 1 person viewing this thread.
These 105 comments are related to the topic:
What's the problem about bots?
What's the problem about bots?
sort by:
Show
per page
Replies: 105
|
By:
aye robot, are you finding it harder now with the declining liquidity in-running?
|
|
By:
are you finding it harder now with the declining liquidity in-running?
It's hard to say - on the one hand my returns in 2016 have been good, better than 2015 which was better than 2014 and I'm optimistic about 2017. On the other hand my operation is getting ever more sophisticated so I can't say what my returns would have been had I not innovated. I had the great good fortune to start in the days before the cross-matcher when it was really easy. When that first came in I had a tough year immediately after but since then I've rebuilt and I now exceed even those heady days. I think if I started now in the way that I did then I wouldn't succeed, but whether that's because of reduced liquidity or because people like me are getting more effective all the time I couldn't say. One of the things about being a long term player is that you build resources like banks of data that new players simply can't access so new entrants now face far tougher competition than I did. It is undeniable that less money is being matched - it's there in Betfair's own data (freely downloadable to all) but even now there's always £100k matched in running on even the cr@ppiest races, and there are circa 12,000 races a year in the UK and Ireland so it's not like there's a shortage of opportunity. It's also hard to say what reduced liquidity really means, what causes it or whether it's good or bad. In a sense I have reduced total liquidity in the past year myself because I'm becoming more accurate and targeted in my betting. As a result I've reduced my transaction count and improved my average profit per bet to make an overall gain. That means I'm giving away less value to the average punter which is bad for punters. On the other hand I'm out competing other botters, if that helps bust out some of the more marginal Bot players that'll certainly mean lower liquidity but on the flip side it'll mean that there's more opportunity to match value bets both for me and for all the manual players. The cause and effect chains of all this are so complex that I don't think anyone can really say what the ultimate effects of a given cause are. In the end I think we all have to be careful what we ask for - I'm not sure I want more liquidity as more efficient markets mean less opportunity to get value. On the other hand I do want people to play because otherwise I can't take their money, and I want them to keep playing so they need reasonably efficient markets to keep them entertained. There must be some sort of Golilocks situation but I don't think anyone knows what that is. |
|
By:
but aye robot, what is your criteria to assess whether it's a value bet or not?
Well I can't tell you that can I... that's my whole gig. Broadly though - it's like I said above, I identify what types of betting behaviors typify over-reactions or panic betting and bet against them. |
|
By:
aye robot, I know in the past you've said that you never "trade out" a position. So what's your opinion of the typical short term trading (open a position and then close the position for a profit or loss, usually pre race) and people trying to use bots to do this? Do you believe it can never really be profitable and is a fool's errand or is it just something you've not been interested in doing?
|
|
By:
Greening out your positions just for the sake of it or because you're scared of the liabilities isn't optimal, which isn't to say that you can't do it and still be profitable.
The starting point for understanding why is that whatever strategy you use you must place value bets, there is simply no combination of bets that returns a profit without showing value. This applies to everyone whether or not they like or understand it and it applies equally to 'traders' automated or otherwise. Now let's say you can somehow identify a signal that indicates that a price will shorten - what you've found is an indicator of value to back. Say you take that back position and get matched, in theory you can green out without any reduction in returns by taking a true value lay bet (zero EV if you prefer that term), but in practice that's not quite true. It's not true because all else being equal your green out bet is more likely to be matched the worse value it is. Therefore if you're not correctly identifying the value of your closing bet - and if its value is not zero or better you're throwing away money by closing out. On the other hand if you can identify a value lay bet then you should take it. The point is that the two bets have nothing to do with each other and should be treated separately if you want optimal returns. There might be other reasons to green out like freeing up funds and depending on your situation it might save you comm (although I doubt that's often true) but in pure betting terms it's not beneficial. |
|
By:
looking at my post above I didn't express something well, this is what I should have said regarding closing out a value back position with a lay bet:
Looked at simplistically if you bet randomly into the exchange you get true value (zero EV) and therefore break even pre-comm and just take the marginal loss of you comm rate. This suggests that if you take a series of positive value back positions followed by lays made without consideration of value then your lays should break even, your backs will win and you will be in the same position as you would have been if you only backed. However in practice this isn't true because not all bets are matched and when you bet without considering value your poor value bets will be matched more often than your good ones. Your good value back bets either matched (good) or didn't (neutral) but either way you didn't risk throwing away money on poor value bets. Or you can think about the same thing this way: You wanted to open a back position so you waited for a positive signal and put in a back order at positive value. Maybe it matches or maybe not, but you can take it or leave it because if it doesn't match that's fine you just don't make any money, therefore you can set the price and be sure only to bet at good value. Now assume your opening bet matches and you're absolutely determined to close out - that means you go from being able to choose whether or not to bet to being forced to bet even if its value is negative. That means that the average value of your closing bet will be negative (for reasons described above) because you don't have the option to just not bet. |
|
By:
Interesting stuff, aye robot. I'm not a botter, don't touch horse racing and no inplay for us Aussies, but profitably do generally the same as you.
I will take issue with one thing you said, there is simply no combination of bets that returns a profit without showing value. Many people reading that will think each and every bet has to be a value bet, which is untrue. You can make a poor value bet if you anticipate making another bet, which may also be a poor value bet, if they net a profit. I do this all the time, based on what I call the psychology of markets, which you allude to. * Although the term 'value' is a rather more slippery concept than many assume. |
|
By:
i suppose the only people bot users like yourself are winning money off are other bot users and those who bet manually inplay
and anyone who does inplay manually is going to be up against it imo, especially if you don't have fast pics don't think bot users make any difference to someone who just wants to take a position on the market beforehand. Like me  |
|
By:
Very insightful. Thanks aye robot.
|
|
By:
Some smashing gen from aye robot - thanks for that.
Interesting that Lewisham ranger mentions fast pics in response - it just shows how two diametrically opposed methods can yield results from the same market, fast-pics-user with race-reading and AR with his market-reading. In practice, fast-pics-user can still get hammered if he assesses value incompetently. There used to be an old fella on the horse racing forum who paid high PC and said he seldom won more than £10 a race. His skill was reading the race better than most people and he did it all from slow video. (I'm sure PC was not designed to catch people like him, but BF took his money anyway.) Aye Robot - do you play on the "low-mayhem" I/R markets e.g. golf, cricket? |
|
By:
Aye Robot - do you play on the "low-mayhem" I/R markets e.g. golf, cricket?
No - I'm working on it though. My main focus this year will be on football and tennis but recent changes to the API (the means by which you communicate with Betfairs servers) mean that it's now much easier to get data from many markets simultaneously. That's great for the likes of me because it means I can build really good simulations of markets I have no experience of. I think there might be more mayhem in cricket than first appears. It's certainly less appealing to botters though just because the events take so long and there aren't that many of them. A lot of what makes horse racing good is that there are so many events and they're over quickly. As a rule of thumb I think it takes me about 1000 markets to see my variability shake out. It takes a long time to bet on that many cricket matches but it's only a month's worth of horse races. The other thing is that the delay in horse races is only one second which doesn't give a human any time at all to make a reasoned decision let alone get a cancel order in whereas the 5 or more that you get on other sports is ample time to weigh up quite a lot. |
|
By:
Interesting.
Have you switched to the streaming API? Or are you in the process? |
|
By:
I've started switching but I'm doing it with caution. I've got it all written now but I'm moving my betting over slowly and watching with care.
|
|
By:
Yeah, I was interested, but I don't rely on speed the way you have to, so it's not a priority yet.
|
|
By:
hate to break it to you, but this user aye here you all seem to be in awe of is also all you guys used to complain about
he doesn't care much about sports, doesn't really (have to) gamble but instead relies on bots (enemy #1) and fast data (enemy #2) to pick of others, even using cross-market-matching algorithms (enemy #3) to do that.....there isn't much more to it, even though he tries very hard to convince others and most importantly himself that it's about his super models and the "big, meaty gambles" he, the big gunslinging gambler, takes |
|
By:
I was interested, but I don't rely on speed the way you have to, so it's not a priority yet.
It's really less about the speed and more about being able to get a lot of markets at once efficiently. That plus getting it all sorted out ahead of the time when it may become compulsory. |
|
By:
I kind of agree with you fixed. As I said earlier, I don't think aye robot is a gambler at all.
Problem is can we really complain about bot use on an automated betting website? |
|
By:
The bloke is totally honest about what he does. Just because he manages to do something
that most people can't doesn't make him a wrong 'un. If you could work out how, you'd copy him in a heartbeat. Is he a gambler? Yes, but he puts the odds on his side. He still risks large amounts of money and exposes himself to the risk of a long series of heavy losses. Even a casino can be broken if someone wagers enough and gets lucky enough. Instead of carping, learn from him while he's generous enough to share his knowledge. |
|
By:
He's not a gambler. He's an arbitrager, that's something completely different. He's winning money from the fluctuations in markets, rather than actually taking an opinion on an event.
|
|
By:
Aye Robot, What do you tell new people you meet about what you do for a living if they enquire? Do you struggle to explain it? Do you call yourself a professional gambler?
|
|
By:
What do you tell new people you meet about what you do for a living if they enquire?
Mostly I keep my activities on here completely quiet. I also do something else and I don't want the people from that field to know about this one - also I don't want the security problem of people coming after my computers so I keep it all very low key. When I do tell people about this I explicitly talk about 'gambling' and take care never to use any euphemisms. I started that rule as a discipline for myself. When I first began I was really strict that I would never allow myself any illusions about what I was getting into. I knew that gambling was very problematic for a lot of people and I didn't want to be one of them so I took a lot of care never to be any kind of fantasist. Being straight forward about the terminology was part of that for me. |
|
By:
Do you struggle to explain it?
Yes. There is something about probabilistic thinking that seems to be fundamentally offensive to the human brain. Many people simply can't (or won't) grasp the fundamental concepts. My closest friend has known about what I do since the start and I have tried to explain it to him on numerous occasions but he just doesn't get it. He's not thick, but there is something in his brain that just will not allow him to see the connection between things that seem to be random and things that don't. |
|
By:
I agree, the sucking up to this guy is becoming rather sickening.
Clearly he's very well skilled at what he does, however he's making money out of people who have a passion for gambling while he knows next to nothing about what he's betting on. good luck to him but you can count me out of the general love-in. |
|
By:
Lewisham - what do you think would happen to the money if I didn't win it? It would just be sucked up in commission. At the level of the individual bet I'm taking my counter-party's money (or else he takes mine) but when I withdraw it's Betfair's money that I'm taking because that's where it would go if it didn't go to me.
For me to bet there has to be a gap between what's being traded and true value, I make money by narrowing that gap and keeping the difference. That means that I keep the market closer to true than it otherwise would be and in doing so I keep the average punter enjoying his play for longer than he would without me. It's true that I marginally reduce his chance of becoming a net winner but let's face it - he wasn't going to be a winner anyway. Perhaps people like me change his chances of showing an overall profit from 0.1% to 0.09% - is that such a big deal? I'm not here because I want to help the punter but then neither are you. I'm here to make money, but in doing so I form part of the total exchange ecosystem that we all depend on. |
|
By:
aye robot is just part of the evolution process of Betfair. Of those who join Betfair some are capable of taking advantage of the opportunities that it offers but most aren't so go the way of the Dinosaur.
No point in complaining either adapt or die. |
|
By:
Arbitrage is something else.
Yep. Arbitrage is misunderstood on here. I can't recall ever seeing a genuine reference to arbitrage, which is broadly speaking taking opposing positions in 2 or more separate markets. I do arbitrage and have a fairly big one ongoing at the moment. There is something about probabilistic thinking that seems to be fundamentally offensive to the human brain. I've long maintained the only thing of any practical use formal education taught me was a basic understanding of probability. I don't bot, but I know enough to know what can be done with them to avoid sports and markets that are conducive to botting. |