POLL: Which team is better-off after a 0-0 in the first leg of a two-leg tie?

[montpelier]

but it should be pretty obvious that man city would be more likely to draw away at a schalke than at home.

I don't have a dog in this fight except to tend slightly towards there being a data discrepancy.

It might be possible to argue though, that any team strength factor is mitigated somewhat by the 1st legged nature of the encounter.

So not all that significant, at the end of the day, after we carry on & look at something else, which I think was wins on away goals not being supremely statistically higher wasn't it.

Our data sample has an allegedly higher % of stronger away teams - but we are still seeing an outcome that goes against this.

Is that right? edit - No, it's the 2nd leg home teams that are allegedly stronger than they averagely ought to be.

Great thread either way, mind.

 

[Pogue Mahone]

Just wanted to prove how stupid your you'd be wrong was since there's clearly tons of evidence that he's not.

I honestly don't know what point you're making here. Sure, you can argue that a 0-0 result in the first leg proves that teams are evenly matched. You can argue whatever the hell you want but the result in a single 90 minute game of football is a ridiculously small sample to draw any kind of meaningful conclusion. Burnley played in Old Trafford recently and were 2 nil up with 5 minutes left. Does this mean Manchester United and Burnley are evenly matched?

The question in this thread is whether, in general, a 0-0 result in a CL knockout tie favours the team playing at home or away in the second leg? All the available evidence (and common fecking sense!) tells us that yes, the team playing at home have the edge here.

Yet somehow we've ended up a bunch of people who are so befuddled by the away goal rule (plus Liverpool fans) that they're willing to bend logic all out of shape and argue to the contrary. Fill your boots but reality is against you on this one.

 

[noodlehair]

Bizarre logic in this thread.

It does depend on the teams but as a general rule I would have thought it's fairly obvious it favours the home side in the second leg.

In European competition especially, if you consider yourself a contender to win then you would be confident of beating an opponent at home. If you can't beat them in your own stadium, you are obviously less likely to beat them anywhere else.

Also, the first leg ending 0-0 indicates that the opposition were unable to beat you, despite having a home advantage, which would indicate they are not the better team.

Obviously there are other factors but then you are getting into speculating based on individual circumstances and predicting an outcome of a one of game, and the reason you're allowed to bet on football is because this is generally impossible to do accurately.

If you look at the Bayern vs Liverpool tie as an example, you have a bunch of people here claiming it suits Liverpool based on, I don't know, playing style? But this ignores that Liverpool have lost their last 5 away ties in Europe (if you include the final last year), so are clearly a stronger team at home. When you look at the United vs Sevilla tie last season, no one would have told you before the game that Sevilla were favurites because they drew the first leg 0-0.

 

[RobinLFC]

I honestly don't know what point you're making here. Sure, you can argue that a 0-0 result in the first leg proves that teams are evenly matched. You can argue whatever the hell you want but the result in a single 90 minute game of football is a ridiculously small sample to draw any kind of meaningful conclusion. Burnley played in Old Trafford recently and were 2 nil up with 5 minutes left. Does this mean Manchester United and Burnley are evenly matched?

The question in this thread is whether, in general, a 0-0 result in a CL knockout tie favours the team playing at home or away in the second leg? All the available evidence (and common fecking sense!) tells us that yes, the team playing at home have the edge here.

Yet somehow we've ended up a bunch of people who are so befuddled by the away goal rule (plus Liverpool fans) that they're willing to bend logic all out of shape and argue to the contrary. Fill your boots but reality is against you on this one.

Just to be clear, I've never argued against the bolded part.

There's only one correct answer to the theoretical question in the thread title, and it's the away team. But the question stops there and is nothing more than just that - theoretical. It seems to me like a lot of people are confusing that with the question "who has the best odds to progress after a 0-0 in the first leg", because the answer to that depends on a dozen other criteria, the quality of the teams involved obviously being the main one.

I believe my first reply in this thread was that I'd back Liverpool to progress at Anfield against almost all teams in the world after a 0-0 away from home.

 

[Pagh Wraith]

The question in this thread is whether, in general, a 0-0 result in a CL knockout tie favours the team playing at home or away in the second leg? All the available evidence (and common fecking sense!) tells us that yes, the team playing at home have the edge here.

No. All available evidence, empirical data and common sense says that a 0-0 has no effect on the probabilities before the first leg. This has been established and proven and it is mind-boggling that this thread is still going. A 50-50 tie is still a 50-50 tie after a goalless first leg making the 0-0 a unique scoreline in that regard. Every other scoreline alters the probabilities significantly.

 

[711]

A 0-0 has no effect on the probabilities before the first leg? This thread's completely bonkers.

 

[Pagh Wraith]

A 0-0 has no effect on the probabilities before the first leg? This thread's completely bonkers.

On average, yes. The bigger the gap in quality between the two sides the bigger the effect. (see the slight drift in Barca's odds). With two evenly matched sides, it has NO effect. We have explained it multiple times now.

 

[Siorac]

A 0-0 has no effect on the probabilities before the first leg? This thread's completely bonkers.

Yes, Pagh should be mostly just ignored. He doesn't really understand the word "evidence", confuses it with "I think so". His 50-50 theory (that is, in the case of evenly matched teams a 0-0 in the first leg does not change the odds of progressing) is an acceptable hypothesis but it hasn't been backed with any sort of data yet.

 

[RK]

To write what others have been saying another way:

If I told you that a game finished 0-0 at a neutral venue, and gave no other information, you could not conclude that either team is superior.

If I told you that a game finished 0-0 at a team's home ground, and gave no other information (other than some home advantage exists in the sport), you must conclude that it is more likely that the away team is superior.

This means any further analysis over a sample of 0-0s at a non-neutral venue must control for these effects.
I think the best way of doing this in football is looking at historical bookmaker odds, but it's not a simple task.

The question in this thread is whether, in general, a 0-0 result in a CL knockout tie favours the team playing at home or away in the second leg? All the available evidence (and common fecking sense!) tells us that yes, the team playing at home have the edge here.

I think that's a valid way of interpreting the question but a bit pointless.
It's just a less extreme way of asking "Does a 0-5 away win in a CL knockout tie mean the home side in the second leg is more likely to win that second leg match?" The data would come back with 75% home wins in the second leg because in reality the sample comprises good teams smashing a bad teams.

Similar to how a sample of 0-0s is more likely to be a superior away team drawing with an inferior home team.

 

[BarcaSpurs]

t's best just to admit it.

The list of bookmakers that according to you are not worth their salt has the likes of William Hill, Ladbrokes, Coral, Betfair, Paddy Power, Skybet and more on it. I'm pretty sure their business models aren't based on simply copying the odds set by marathon and 365.

Correct they copy Pinnacle and Betfair Exchange.
Models update with new information causing prices to move, then whoever is sticking out will shift their odds, then whoever is now sticking out will shift their odds etc.
If we see an unusually large bet come in on something all we do is check the market and see if we're in line, we don't adjust unless theres something wrong with the odds.

I already told you this is my job so I'm not sure why you're so insistent on telling me I don't know how it works.

 

[montpelier]

when does ''before the first leg'' become a factor?

 

[711]

Yes, Pagh should be mostly just ignored. He doesn't really understand the word "evidence", confuses it with "I think so". His 50-50 theory (that is, in the case of evenly matched teams a 0-0 in the first leg does not change the odds of progressing) is an acceptable hypothesis but it hasn't been backed with any sort of data yet.

Well right or wrong that at least makes sense if he means the odds of progressing, but what he actually said was 'no effect on the probabilities before the first leg'. Maybe I'm being deliberately obtuse now, it's just every other post being slightly loopy has got to me.

 

[711]

Correct they copy Pinnacle and Betfair Exchange.
Models update with new information causing prices to move, then whoever is sticking out will shift their odds, then whoever is now sticking out will shift their odds etc.
If we see an unusually large bet come in on something all we do is check the market and see if we're in line, we don't adjust unless theres something wrong with the odds.

I already told you this is my job so I'm not sure why you're so insistent on telling me I don't know how it works.

Good luck in your career.

 

[montpelier]

a 0-0 in the first leg does not change the odds of progressing

yes, but this is because of the away goals rule though, isn't it?

you're trying to confuse us with mistaken simplification aren't you?

 

[RK]

His 50-50 theory (that is, in the case of evenly matched teams a 0-0 in the first leg does not change the odds of progressing) is an acceptable hypothesis but it hasn't been backed with any sort of data yet.

In this case, wouldn't his theory be the null hypothesis, thus the burden of statistical proof rests with other theories?

For what it's worth I'm mostly in agreement with Pagh and BarcaSpurs.

 

[montpelier]

Shall we ask for a poll?

 

[Siorac]

yes, but this is because of the away goals rule though, isn't it?

you're trying to confuse us with mistaken simplification aren't you?

Maybe I shouldn't use the word "odds" and should have stuck to "probabilities", as I'm not interested in betting odds but in actual real world outcomes.

And the question is that away goals or not, does a 0-0 in the first leg between two evenly matched teams increase or decrease the chances of either one progressing?

 

[Danny Roberts]

edit - I'll shut up actually.

 

[RobinLFC]

Maybe I shouldn't use the word "odds" and should have stuck to "probabilities", as I'm not interested in betting odds but in actual real world outcomes.

And the question is that away goals or not, does a 0-0 in the first leg between two evenly matched teams increase or decrease the chances of either one progressing?

In isolation? Yes.

In reality? It simply depends on multiple factors, e.g. for the Liverpool game, Kimmich suspension and Van Dijk back could offset the home-field advantage for Bayern and you're back where you started. If you only know who's playing at home, you should say you have insufficient data to come to a conclusion. If you have to pick one gun to your head, always the home team.

 

[Siorac]

In this case, wouldn't his theory be the null hypothesis, thus the burden of statistical proof rests with other theories?

For what it's worth I'm mostly in agreement with Pagh and BarcaSpurs.

Well, to make that a null hypothesis it would have to sound something like this: "there is no relationship between the first leg of a knockout tie finishing 0-0 and the eventual winner of the tie when teams are evenly matched". And it's a hypothesis that is not THAT difficult to test, assuming we can come to an accepted definition of what constitutes "evenly matched" and compile a reasonable sample size.

But frankly, all this statistical analysis is a little tedious and probably misses the point. In the real world of football, a 0-0 draw away from home in the first leg is generally a decent result: teams that can achieve it go through way more often than not. The reasons why that is so are actually not that relevant: even if it's because the away teams in the first leg are generally superior teams, well, with a 0-0 it's likely that they can make that superiority tell.

 

[Pagh Wraith]

Well right or wrong that at least makes sense if he means the odds of progressing, but what he actually said was 'no effect on the probabilities before the first leg'. Maybe I'm being deliberately obtuse now, it's just every other post being slightly loopy has got to me.

Probabilities before the first leg means the odds of progressing before kick-off. Sorry if I worded this badly. SBOBET offered 1.99 for Liverpool to advance before the first leg and 1.96 now. So the mulit-million pound, highly efficient betting markets neither agree that the first leg result has altered the chances significantly nor that this was a great result for Bayern. And if you go back through historical data you will see that this is the case for every evenly matched tie.

 

[711]

Shall we ask for a poll?

@Damien ?

Probabilities before the first leg means the odds of progressing before kick-off. Sorry if I worded this badly.

No, I should have realised, ignore me.

 

[BarcaSpurs]

Well, to make that a null hypothesis it would have to sound something like this: "there is no relationship between the first leg of a knockout tie finishing 0-0 and the eventual winner of the tie when teams are evenly matched". And it's a hypothesis that is not THAT difficult to test, assuming we can come to an accepted definition of what constitutes "evenly matched" and compile a reasonable sample size.

No, again that isn't how hypothesis testing works and isn't at all what either of us are arguing.
We aren't saying there is no relationship, infact, as I said earlier, that'd be almost impossible. What we are saying is that there is insufficient evidence to prove the relationship is favourable towards the away team in the first leg. Likewise there is no evidence suggesting the home team is favoured with a 0-0.

The 2 outcomes are we either reject the null hypothesis in favour of the proposed hypothesis, or there is insufficient evidence to prove the proposed hypothesis. at no point do we state that the null hypothesis is true.

If I roll a dice 100 times and get 25 6's is the Dice biased? probably not, but im also not saying it's definitely not biased either.

 

[Spiersey]

You quite clearly said bookmakers don't change their odds according to the weight of money placed, and you were wrong. Now you've changed to they 'adjust odds to keep in line with the market' - yes, a market that moves according to the weight of money placed. If relevant factors alter then they will adjust their odds immediately to pre-empt liabilities, for example if a star player was declared injured, but if the money didn't move as expected they would revert. You can think of something else to throw in if you want but I'll only bring you back to your original statement being wrong, sometimes it's best just to admit it.

The list of bookmakers that according to you are not worth their salt has the likes of William Hill, Ladbrokes, Coral, Betfair, Paddy Power, Skybet and more on it. I'm pretty sure their business models aren't based on simply copying the odds set by marathon and 365.

Every bookie in Britains business model is based off following Asian bookmakers (Pinnacle, Singbet, ISN, SBO) for 80% of games and the other 20% of obscure games they all follow 365. If a star player was injured in a mainstream game, no bookie would manually adjust the odds as they'd run the risk of being massively wrong, they would just wait for any movements from money in Asia. If a big player is out the money definitely comes, Barca Asian handicap price drops 0.3 whenever Messi is subbed on/off which takes 6 figures over a minute or two.

 

[RK]

Well, to make that a null hypothesis it would have to sound something like this: "there is no relationship between the first leg of a knockout tie finishing 0-0 and the eventual winner of the tie when teams are evenly matched". And it's a hypothesis that is not THAT difficult to test, assuming we can come to an accepted definition of what constitutes "evenly matched" and compile a reasonable sample size.

I'd be interested in this, but I still think we'd have sample size issues if just counting even matches. Think a better route would be to include all matches and fit some model to it. E.g. the midpoint between "teams unevenly matched by +z%" and "teams unevenly matched by -z%" would have some value. But like you say we'd still need to agree on some rating system which isn't easy. I'd suggest historical bookmaker odds which aren't easy to obtain publicly for To Qualify, or would have to be adjusted for home advantage if using Full Time Result.

Obviously this is getting a bit statsy for a football thread now but I like it.

But frankly, all this statistical analysis is a little tedious and probably misses the point. In the real world of football, a 0-0 draw away from home in the first leg is generally a decent result: teams that can achieve it go through way more often than not. The reasons why that is so are actually not that relevant: even if it's because the away teams in the first leg are generally superior teams, well, with a 0-0 it's likely that they can make that superiority tell.

This is true but, as I went into more detail earlier, I don't think it's the interesting take on the question because the answer is obvious.

 

[711]

Every bookie in Britains business model is based off following Asian bookmakers (Pinnacle, Singbet, ISN, SBO) for 80% of games and the other 20% of obscure games they all follow 365. If a star player was injured in a mainstream game, no bookie would manually adjust the odds as they'd run the risk of being massively wrong, they would just wait for any movements from money in Asia. If a big player is out the money definitely comes, Barca Asian handicap price drops 0.3 whenever Messi is subbed on/off which takes 6 figures over a minute or two.

Thanks for that, I'm willing to learn that the Asian market, due to volume no doubt, is far more influential than it was in the days I was involved. The dispute I was having though centred on my assertion that it was weight of money that mostly determined, and changed, the odds on offer. I stand corrected that most of that weight now comes from Asia, but retain the assertion.

 

[montpelier]

And the question is that away goals or not, does a 0-0 in the first leg between two evenly matched teams increase or decrease the chances of either one progressing?

When you were arguing though...

And I don't really know what to say here. The rest of us would have agreed with what I've quoted. But previous to that was talk about a sample of 2nd legs where the first leg had ended 0-0. It was then suggested that this was a flawed sample, yes. I don't know who by, it might even have been you.

At which point, the argument carried on, but it seemed to me that most people were still in agreement with the common-sense idea that 0-0 result favoueed the team playing the 2nd leg away especially if they were possibly stronger too - as would be self-evident because of the result of the 1st leg.

Someone talked about a sample of 12 0-0 draws compared to a random sample of 12 0-0 draws & there being a possible flaws in the methodology of the stats-bashing because of this. The stats seemed fairly inconclusive of taking us away from the common-sense view OF US ALL anyway,

I'd just like to ask therefore, honestly, what is it that you think you've been arguing about or is it all @Pagh Wraith 's fault?

Feel free to accuse me of being confused because I am, .

 

[Siorac]

No, again that isn't how hypothesis testing works and isn't at all what either of us are arguing.
We aren't saying there is no relationship, infact, as I said earlier, that'd be almost impossible. What we are saying is that there is insufficient evidence to prove the relationship is favourable towards the away team in the first leg. Likewise there is no evidence suggesting the home team is favoured with a 0-0.

The 2 outcomes are we either reject the null hypothesis in favour of the proposed hypothesis, or there is insufficient evidence to prove the proposed hypothesis. at no point do we state that the null hypothesis is true.

If I roll a dice 100 times and get 25 6's is the Dice biased? probably not, but im also not saying it's definitely not biased either.

That might be what YOU are arguing but definitely not what Pagh is. He says that 0-0 does not alter the chances in evenly matched ties and that the evidence proves it. Read back his posts. He's not saying there is insufficient evidence to prove a relationship. That is indeed a different hypothesis but again, that's not what he is saying.

And frankly, there is very little real value in your position either. It adds nothing to the discussion in the end. Let's say I claim that a 2-0 home win in the first leg clearly favours the team that played home in that first leg. Can we find sufficient evidence that it's true? We would have to rule out all cases when the home team was clearly superior to begin with and then check the results. Is there a big enough sample size of evenly matched ties where the home team won the first leg 2-0? Probably not: after all, in the European Cup/CL between 1979 and 2008, there were 429 0-0 draws and 413 2-0 home wins in first legs of two-legged ties so the sample size is similar. Which means there is probably insufficient evidence overall once we control for team quality.

Does common fecking sense tell us that a 2-0 home win is a very good result? Yes it does.

 

[montpelier]

Oh Christ, there's a fecking poll now, .

 

[Spiersey]

Thanks for that, I'm willing to learn that the Asian market, due to volume no doubt, is far more influential than it was in the days I was involved. The dispute I was having though centred on my assertion that it was weight of money that mostly determined, and changed, the odds on offer. I stand corrected that most of that weight now comes from Asia, but retain the assertion.

You're correct that the money determines the odds now but its only money in the asian market that will effect the odds, Euro bookies will disregard any big bet as their max stake is 20k (and anyone profitable would be banned anyway) whilst Asian bookies will lay much bigger odds so it's silly to make any manual adjustments for anything staked there. A lot of bookies now barely have any traders, think it's betvictor that only have a couple left. Theres basically no need for them for some books as they don't price stuff and price movements and adjustments are generally done by computer/algorithm

 

[Siorac]

When you were arguing though...

And I don't really know what to say here. The rest of us would have agreed with what I've quoted. But previous to that was talk about a sample of 2nd legs where the first leg had ended 0-0. It was then suggested that this was a flawed sample, yes. I don't know who by, it might even have been you.

At which point, the argument carried on, but it seemed to me that most people were still in agreement with the common-sense idea that 0-0 result favoueed the team playing the 2nd leg away especially if they were possibly stronger too - as would be self-evident because of the result of the 1st leg.

Someone talked about a sample of 12 0-0 draws compared to a random sample of 12 0-0 draws & there being a possible flaws in the methodology of the stats-bashing because of this. The stats seemed fairly inconclusive of taking us away from the common-sense view OF US ALL anyway,

I'd just like to ask therefore, honestly, what is it that you think you've been arguing about or is it all @Pagh Wraith 's fault?

Feel free to accuse me of being confused because I am, .

My initial point was that common sense and the numbers agree: a 0-0 draw is a decent result away from home in a first leg.

Others pointed out that there are mitigating factors that can skew the numbers and a 0-0 draw may or may not favour either team when it's evenly matched teams. There seem to be two schools of thought here: one is that we simply cannot know because there isn't enough evidence and another that says a 0-0 draw does not change anything (again, when we're talking about evenly matched teams).

That's about it.

 

[711]

You're correct that the money determines the odds now but its only money in the asian market that will effect the odds, Euro bookies will disregard any big bet as their max stake is 20k (and anyone profitable would be banned anyway) whilst Asian bookies will lay much bigger odds so it's silly to make any manual adjustments for anything staked there. A lot of bookies now barely have any traders, think it's betvictor that only have a couple left. Theres basically no need for them for some books as they don't price stuff and price movements and adjustments are generally done by computer/algorithm

Thanks, I see that. Not the same, but I miss the days when you could bet in person at racecourses, with an edge from not having to pay tax, and watch the bookies making their own books, with a lot of copying no doubt, and try and gain from the odds as they swung, sometimes quite wildly, up and down. Nowadays the bookies are still there, all lined up, but with identical odds straight off the internet. Another piece of the 'old world' gone.

 

[OutlawGER]

It really depends on the situation.

Usually i'd say that the home team is in favor, they managed to get a draw away and now can go ahead and win at home. Most teams are stronger at home, this is a fact. But sometimes different scenarios are playing out... For example Bayern-Liverpool. Bayern might go for the kill in this game at home and play much more offensive. This could play exactly into Liverpools cards because this time they get more space to hit Bayern on the counter, and this is what every Klopp team likes to do in perfection. And if they hit them once, Bayern need to go even more offensive. We have seen something like that against Real Madrid in 2014 where Bayern lost 0:4 or so at home because they opened up after Ramos scored from a simple corner and a freekick by header. Match luck killed that tie very early. So the word match luck heavily favors the away team, because away goals are much more relevant.

But i still say that the home team is in favor. Simply because teams are usually stronger at home. You don't go to Nou Camp or Bernabeau for example and expect to beat the home team. You just don't. Also there is a potential extra time which means you could have more playtime at home.

 

[montpelier]

My initial point was that common sense and the numbers agree: a 0-0 draw is a decent result away from home in a first leg.

Others pointed out that there are mitigating factors that can skew the numbers and a 0-0 draw may or may not favour either team when it's evenly matched teams. There seem to be two schools of thought here: one is that we simply cannot know because there isn't enough evidence and another that says a 0-0 draw does not change anything (again, when we're talking about evenly matched teams).

That's about it.

OK, that's great. Barring interventions, we should all be back on the same page now.

 

[Kapardin]

Always the home team in my view. It removes the away goals headache for them and transfers it to the opponent. Personally I'd prefer to get the home game out of the way first for this reason.

Assume it won't matter if away goals rule is scrapped.

 

[montpelier]

But apparently not all on the same page, I'm amazed.

2/3 of people aren't convinced that drawing 0-0 away - having a slightly better than average chance of being the stronger team, gets you into pole position

Presumably because of the pesky away goals rule mainly. And whatever the 'other factors' that don't average out might be.

 

[Siorac]

But apparently not all on the same page, I'm amazed.

2/3 of people aren't convinced that drawing 0-0 away - having a slightly better than average chance of being the stronger team, gets you into pole position

Presumably because of the pesky away goals rule mainly. And whatever the 'other factors' that don't average out might be.

Well, the "impossible to say" argument is respectable, I think. That has some merit.

Saying the first leg's home team is better off after a 0-0 is completely bonkers though.

 

[adexkola]

I may have missed this, but on average, what are the odds the away side secures a win or 0-0 draw at home?

 

[Pogue Mahone]

Probabilities before the first leg means the odds of progressing before kick-off. Sorry if I worded this badly. SBOBET offered 1.99 for Liverpool to advance before the first leg and 1.96 now. So the mulit-million pound, highly efficient betting markets neither agrees that the first leg result has altered the chances significantly nor that this was a great result for Bayern. And if go back through historical data you will see that this is the case for every evenly matched tie.

I can't believe you're seriously using the betting odds from one game to back up your argument?

Liverpool were the CL finalists last season, have lost just one game in the league so far and played Bayern without, arguably, their most important player this season. Of course the bookies (and pundits) will fancy their chances of beating a Bayern team that's fallen below their usual standards.

None of that has any bearing on whether or not the result favours the home team in the second leg.

 

[noodlehair]

Just to be clear, I've never argued against the bolded part.

There's only one correct answer to the theoretical question in the thread title, and it's the away team. But the question stops there and is nothing more than just that - theoretical. It seems to me like a lot of people are confusing that with the question "who has the best odds to progress after a 0-0 in the first leg", because the answer to that depends on a dozen other criteria, the quality of the teams involved obviously being the main one.

I believe my first reply in this thread was that I'd back Liverpool to progress at Anfield against almost all teams in the world after a 0-0 away from home.

I'm not sure what this theoretical argument is based on.

You are arguing that the away team has an advantage because they can win with a score draw, but this ignores the obvious home advantage as well as the fact the away team were unable to score a goal against their opponent even when it was them who had home advantage.

These aren't speculative factors, they are actual ones. As you just said, you'd back Liverpool to progress if they were the home team. Home advantage is an extremely big factor in European ties. The whole concept of the away goal rule is to try and account for this.