Arbusers

If I understood it correctly value betting is about chosing a certain amount of bettingsites, that are used as a base to calculate true odds based on the implied odds. At the end you multiply the odds with 1/1.05 if the margin is 5 percent, to get the an approximation for the true odds. Then since you want to account for variance and mistake of estimation of true odds, you give yourself rules. If the odds q given by a certain cite are p lower than the true odds you bet. P varies based on the size of the true odds. If the true odds are 0.1 p needs to be chosen differently than if the true odds are 0.8 to account for variance. Are these the basics?

As far as I am concered it works on a kind of the odd one out bias. So if some online site is the odd one out you assume they are full of shit instead e.g having insider information. Does this scheme work with high enough p practically. How much time is spent comparing odds.

Let's not overcomplicate things. Value betting is when your odds are higher than the real probability of an outcome.
Make no sense to focus just on certain bookmakers. Yes, they will give you hints about the real price, but this is not enough to identify value. Read some more in the forum. The topic is huge.

How do people estimate probabilities accurately. I was interested in this. Which posts can I read. I am having difficulties trying to learn. In the stickied post there was talk anout advanced statistical methods. What this is, is a very simple way to try to abuse outliers from the market.

Since botuse is recommended in this forum I was wondering how the algorithm behind it works. Botuse for arbitrage betting is clear for me but if you use bots for value betting you need some way to get an estimation of true odds.

There are no true odds as there is no true probability. Nature is deterministic. So theoretically if you know all the causes and influences of an event and its environment the probability is 100% for something to happen.
Putting that aside what you are talking about is total market expected probability at a particular time. So you are looking for average market price. Even that we do not know for sure. Thats because even if you scan all the world bookmakers and make the average odds, it is not accurate since you dont know the sum of total stake for each odds on each bookie. What you are looking in the scanners is like a steam that you follow because it indicates that a bookmaker has a probability to be outpriced. So what you are really betting this way is a probability of a probability of an outcome of an event.
Thats why value betting is complicated.
Will you make money betting this way ? Most probably yes, but not for sure. And the odds to make money are way lower than most people here and especially the alert services would say. Exactly because of the logic explained in the post.

Hans müller wrote: ↑

Thu Dec 12, 2024 11:42 am

How do people estimate probabilities accurately. I was interested in this.
...
need some way to get an estimation of true odds.

In my opinion, it is always much preferable to calculate your own fair odds, reflectin' the probabilities in any market hypothesis.

You can choose whether to:

(1) Purchase and use an external Benchmark/Software of any kind - a very fast and convenient solution, but not the best, as no one has an interest in selling you (as an individual/ entity/ rival in the end) the instrumentation as precise as an atomic clock; I suspect they'll sell you the so-called export version of their product;

or

(2) Develop your own Creation from scratch, that will:
- will fit your needs exactly;
- will work with maximum prediction accuracy that you can adjust empirically, but it will never be perfect, the errors are inevitable, thus the errors are your best teachers in order to try to improve the next version of the Model;
- will be used only by you;
- will give you the satisfaction of constantly trying to improve your Formulas/ Models/ Triggers/ Forecasts over Time, by relying only on your own Mind/ Brainstorming creative process, that's more than awesome!

Provided that you choose (2), the truth is out there:

-Step one, the mandatory Philosophical background:

plato.stanford.edu/entries/scientific-method/

-Step two, formulating your hypotheses, proving something, in the process of creating your own Model, tune-ups or something else, that's a random link for an example of the algorithm:

microbenotes.com/scientific-method/#steps-of-scientific-method


Hope that helps.

I think many people overcomplicate things. I am still making around 8% yield per bet by doing the most simple thing - comparing soft against pinny. What you need is to set the right filters like liquidity, make sure your feed is accurate (you will be surprised how many mistakes alert services or even bookie sites themselves make) and if you do live betting then make sure you take into account "danger zones" and avoid those value traps.

While strategy is very simple implementing it at scale is very difficult. Doing simple things like pre-match arbing will get you limited super fast. But to do advanced things you need either a lot of experience or good tech (mainstream alert services are definitely not something I would consider a good tech). Also it is important to be different and find your niche. If you do the same as thousands other steam chasers then your accounts will get limited really fast. So in other words it is very simple and extremely hard at the same time