Exploring Value Betting in Football Match Odds – Exit Strategies

[TupleVision]

I've recently been focusing on value betting rather than traditional trading, aiming to identify genuine value in the markets. I've found Peter Webb's articles and an older video on predicting draws particularly insightful. Additionally, I came across a great post by Derek (apologies, I can't recall the full username) on using Poisson distribution for football value bets. While Poisson has its limitations as a model for sports betting, that's a discussion for another thread.

To aid my analysis, I've developed a Python-based GUI tool that helps identify potential value opportunities for draws. It utilizes a database in the background, storing key metrics alongside standard football statistics such as average goals for home and away teams. Some aspects are proprietary, but the core idea is to leverage historical data for better decision-making.

At this stage, I’m primarily focusing on laying the draw, as I suspect that draw odds are often overstated due to how bookmakers price home and away odds.

Seeking Advice on Exit Strategies
I'm looking for insights from those experienced in football value betting regarding exit strategies. My current dilemma:

Exit at 70 minutes – This is based on an observation that around this time, the potential loss is roughly 50% of the maximum exposure. The idea here is to smooth out the profit and loss variance over time.
Let the bets run – Given that these are pure value bets, the argument is that closing positions early could dilute long-term expected value. However, due to the relatively low number of qualifying matches per day, gathering a large enough sample size to validate the approach is proving challenging.
Would love to hear from anyone with experience in value betting on football match odds—what exit strategies have worked for you? Any guidance or suggestions would be greatly appreciated!

[Screenshot below is random values I've entered, the edge is never that big!]

[Kai]

Seeking Advice on Exit Strategies
I'm looking for insights from those experienced in football value betting regarding exit strategies.

Are you familiar with the optimal stop theory? https://www.youtube.com/shorts/m2uJUcDoN3M

[ShaunWhite]

Value betting doesn't have an obvious exit because you've no mechanism for assertaining value. Any +ev you're getting on the first bet can easily be lost on the second.

Entries/exits.... Same logic for both, just reversed.

[TupleVision]

Hi Kai, I've seen your posts quite a lot on football threads and always seek your replies to other poster's questions.

I'm vaguely familiar with optimal stop theory but looks like I have a bit of bedtime reading to do tonight. Certainly food for thought, I think what you're saying is, if I can calculate if the cashout at 70mins is greater than lets say 85% of potential profit then exit. Example as per a quick table I've made

[TupleVision]

Value betting doesn't have an obvious exit because you've no mechanism for assertaining value. Any +ev you're getting on the first bet can easily be lost on the second.

Entries/exits.... Same logic for both, just reversed.

Thanks Shaun, are you effectively saying I'm trying to mix trading with betting?

[ShaunWhite]

Thanks Shaun, are you effectively saying I'm trying to mix trading with betting?

Yeah, and I just think you've put a lot of effort into the first bet, and it'll maybe take the same effort on 2nd (the hedge bet). But if you crack that one then you'll have a mechanism to assess the price at any time, and that's the holy grail.

But... Given the efficiency of football markets a hedge at any point, over time, shouldn't be too harmful. As long as you're offering not taking. Make sure you're collecting all the data you can so you can build a bigger sample without spending money.

[TupleVision]

Thanks Shaun, are you effectively saying I'm trying to mix trading with betting?

Yeah, and I just think you've put a lot of effort into the first bet, and it'll maybe take the same effort on 2nd (the hedge bet). But if you crack that one then you'll have a mechanism to assess the price at any time, and that's the holy grail.

But... Given the efficiency of football markets a hedge at any point, over time, shouldn't be too harmful. As long as you're offering not taking. Make sure you're collecting all the data you can so you can build a bigger sample without spending money.

Ok got it, thanks, this makes sense. I've had a super quick blast at throwing something together that I can build on over the next few days, looking back I can't believe I was trying to decide if I cashout on 70 mins or leave it as if they were the only two choices!

[jamesedwards]

If draw odds are overstated you should be backing, not laying.

Blindly exiting is just throwing money away. If your bank can afford the swings then let it ride. If it can't then just stake less up front.

[TupleVision]

If draw odds are overstated you should be backing, not laying.

Blindly exiting is just throwing money away. If your bank can afford the swings then let it ride. If it can't then just stake less up front.

Wait hang on, is this my terrible use of English or is it a fundamental flaw! When I say if draws are "overstated" I mean in some circumstances and not very often the odds on the draw are too low for what I have modelled, therefore I want to lay this right?

Totally agree with blindly exiting now, it was a foolish idea to begin with!

[jamesedwards]

If draw odds are overstated you should be backing, not laying.

Blindly exiting is just throwing money away. If your bank can afford the swings then let it ride. If it can't then just stake less up front.

Wait hang on, is this my terrible use of English or is it a fundamental flaw! When I say if draws are "overstated" I mean in some circumstances and not very often the odds on the draw are too low for what I have modelled, therefore I want to lay this right?

No worries, sounds like just semantics.

To me overstated odds would mean higher than they should be. Overstated chance would mean lower than they should be.

[wearthefoxhat]

Nice work on your python data input model!

FWIW, I've found the entry point to be the most important when value betting. (Shaun/Kai already said something similar). My entry point is from half-time onwards. Gives me a chance to review the key metrics in the first half, compare the current live market odds, decide if there's still any value left. I don't use an exit point, as I need all the minutes of the game (added time too), to give the value elements a chance to work.

I currently use a poisson prediction excel model (zero inflated) to create my own tissue, compare the live market, then estimate book%, home/draw/away. I then have staking calculator that suggests a stake, based on estimated value, (better value. bigger stake..etc).

South v Bourne.png

[TupleVision]

Nice work on your python data input model!

FWIW, I've found the entry point to be the most important when value betting. (Shaun/Kai already said something similar). My entry point is from half-time onwards. Gives me a chance to review the key metrics in the first half, compare the current live market odds, decide if there's still any value left. I don't use an exit point, as I need all the minutes of the game (added time too), to give the value elements a chance to work.

I currently use a poisson prediction excel model (zero inflated) to create my own tissue, compare the live market, then estimate book%, home/draw/away. I then have staking calculator that suggests a stake, based on estimated value, (better value. bigger stake..etc).


South v Bourne.png

Interesting, it sounds like we're both looking at a similar approach. I must admit I really struggle to find any value/edge at all in the more popular leagues, very occasionally a match will pop up but usually only if there's a number of injuries to one team. The market is just too damn efficient.

An in-play approach is a good angle into the game due to the subjective nature of the individual match and fear/greed shaping the market when goals are scored.

ZI Poisson (ZIP) by the way, I love it when you talk dirty

[Kai]

Would love to hear from anyone with experience in value betting on football match odds—what exit strategies have worked for you?

When I LTD it's usually during 2nd half with no exit, maybe I go for some unders too and look for 1-2-3-4 goals then manage that accordingly

Because if one price is out of whack others can be too, I sometimes do similar on the HT market for the 1st half, but wouldn't LTD at KO as that's way too many variables, price swings aren't worth it and wrong teams score

But why so much focus on exits, get your entries right first? When trading you can obsess all you want about perfect exits and maximum profits but no trader can regularly take the perfect exit and extract the maximum profit from each trade because that's impossible, you will have to be content with optimal exits and optimal profit, the priority is to protect your position and your bank first and foremost.

[Kai]

I must admit I really struggle to find any value/edge at all in the more popular leagues, very occasionally a match will pop up but usually only if there's a number of injuries to one team. The market is just too damn efficient.

It's because you're entering at KO when market is supposed to be most efficient, have to enter earlier if you want to avoid that

You're right though, it's way harder to find value in something like Arsenal vs Chelsea than Al-Kebab vs Al-Pacino for example

But bailing at 70 minutes means you're missing on the best and decisive part and hella swings/dopamine, I would try and find a way to profit off late drama in football

For some things are dead cert like death, taxes, and RM comebacks in CL

You work damn fast btw! Small sample size for posts, but whatever you go for I'm pretty confident you nail that

[TupleVision]

I must admit I really struggle to find any value/edge at all in the more popular leagues, very occasionally a match will pop up but usually only if there's a number of injuries to one team. The market is just too damn efficient.

It's because you're entering at KO when market is supposed to be most efficient, have to enter earlier if you want to avoid that

You're right though, it's way harder to find value in something like Arsenal vs Chelsea than Al-Kebab vs Al-Pacino for example

But bailing at 70 minutes means you're missing on the best and decisive part and hella swings/dopamine, I would try and find a way to profit off late drama in football

For some things are dead cert like death, taxes, and RM comebacks in CL

You work damn fast btw! Small sample size for posts, but whatever you go for I'm pretty confident you nail that

Completely agree with this hypothesis, I'm wondering if I now adapt my 'cash out' calculator to model half-time data and see if there's value somewhere else, I think I should be able to put a few more fields in and take out some redundant ones.

I work damn fast as I work from home, my day job is in data analysis which means I'm basically stealing a living from my company right now as I'm spending the best part of 12hrs a day on this and 30mins of paid work, I've just perfected the art of looking busy on Teams calls