Now I'll quite freely admit to not knowing a thing about probability theory, so what I'm about to say could be a load of rubbish (according to that theory, although it makes more sense to me than that theory), but here goes:
Isn't the possibility of any combination of results in two successive matches (i.e. according to your example, Spain/Spain, Spain/Draw, Spain/Scotland, Draw/Spain, etc.) 1/9? OK, so if Spain win the first match, as you say, the chances of them winning the 2nd match would be 1/9. However, the chances of it being a draw would also be 1/9, and so would the chances of a Scotland win. The only reason a second successive Spain win would have a 1/9 chance is because you've already done a large part of the "hard" work in getting the first 1/3 chance right.
For a coin toss, it doesn't matter that you've spun heads 5 times in a row, the chances of the 6th toss being heads are still 1/2. The chances of spinning 6 heads in a row may be 1/64 (I think
), but so are the odds of spinning 5 heads followed by a tail (or 4 heads followed by 2 tails, or alternately spinning heads and tails 3 times each, etc.).
You're right that a certain sequence can't continue forever, it presumably has to change eventually - however, there's no way of knowing when, and you could lose a hell of a lot of money in the mean time waiting for that change to occur.
And in any case, surely all of the above only applies to an event which is consistently the same time after time i.e. tossing a coin, spinning a roulette wheel, etc.? In a football match there are so many variables (opponents, injuries/suspensions, weather conditions, etc. etc.) to take into account as well that go far beyond questions of probability.