Now if there was not much substance to these indications of finetuning in our universe, they would have been brushed off and ignored by those who ponder the universe from a materialistic standpoint. Instead quite the opposite seems to have happened to a large degree. Rather than ignoring these indications of finetuning, many recent thinkers have let these indications drive them to a huge change in thinking: the assumption that there is a multiverse or ensemble of universes.
The idea behind the multiverse is that there are a huge collection of universes, presumably each with a random set of physical characteristics, laws, and physical constants. The person who advances such a concept uses it in an attempt to reduce the “miracle of habitability” of our incredibly improbable universe to make the ultimate “long shot” look like something that is likely to occur at least once. Assume that the chance of a random universe being compatible with the evolution of intelligent beings is only 1 in N, where N is some extremely high number. If we assume that there is a number of universes much greater than N (say N times 1000), than there might actually be a likelihood that some universe would purely by chance meet all of the requirements necessary for intelligent life.
Let us take a close look at this idea of a multiverse, and see whether it holds up to scrutiny, in terms of explaining anything. Before we can do that, we must carefully define some terms and principles.
When talking about probability, mathematicians use the term trials to mean something like an experiment in which different results can occur. A trial may or may not be a formal experiment. It might be something like a hand in a game of cards, a roll of the dice, a particular spin of a roulette wheel, or a particular swing of the bat while a baseball player is at bat. In talking about a possible ensemble of universes, then each different universe would be considered a trial.
Now let me define two important terms.
Term  Definition 
Number of successful trails  The total number of successful trials in a series of random trials 
trial success probability  The chance that a random trial will be successful. When there are a large number of trials, the trial success probability tends to be equivalent to the ratio between the number of trials that are successful and the number of trials that are unsuccessful 
Here are some examples to clarify the use of terms.
Example 1: If one thousand raffle tickets are sold in a raffle, and a winner is selected from a barrel containing the stubs of all of these raffle tickets, then the number of successful trials will be 1 and the trial success probability is 1 in 1000. The number of successful trials is exactly one because there is only one winning ticket chosen. The chance of any ticket holder winning or the trial success probability is 1 in 1000.
Example 2: If I make three bets on some position of a roulette wheel with 38 positions, each time betting on a particular spin of the roulette wheel, then the trial success probability is 1 in 38, because on each spin of the wheel there is one chance in 38 of winning. The number of successful trials depends on chance. It could be as high as 3 if I am very lucky, or it could be as low as 0 if I am unlucky.
Now, what general principles can we state about the relation between the number of successful trials and the trial success probability? They are as follows:
Principle 1: If the trial success probability is zero, the number of successful trials must be zero. This simply means that if there is no chance of one trial being successful, there must be no successful trials.
Principle 2: If the number of successful trials is zero, it does not necessarily mean that the trial success probability is zero. For example, in Example 2 above it is entirely possible that the number of successful trials might be zero, meaning the gambler loses on all three spins of the roulette wheel. But with each spin of the roulette wheel there is 1 chance in 38 of being successful; in other words the trial success probability is 1 in 38.
Principle 3: In some cases increasing the trial success probability will tend to increase the number of successful trials, but an increase in the trial success probability will not always mean an increase in the number of successful trials. For example, let us define a successful trial for a baseball hitter as hitting a home run. A hitter may improve his batting technique, thereby increasing the trial success probability, his chance of hitting a home run. That will tend to increase (over a long enough time) the batter's number of successful trials, or number of home runs, but is not guaranteed to do so during a particular number of trials.
Principle 4: For random trials that do not involve practice at a skill, increasing the number of trials tends to increase the number of successful trials or make it more likely to be at least one, but increasing the number of trials does not increase the trial success probability. For any random process such as the spin of a roulette wheel of the throw of a pair of dice, an increase in the number of trials or attempts may tend to increase the chance of one successful outcome (or increase the number of successful outcomes), but the number of trials or attempts does not tend to increase the chance of success on any random one of the trials, that is the trial success probability. The one exception to this principle is that when the trials involve a human practicing, the chance of success on one trial may actually increase, because of the tendency of a human to improve with practice.
Now let us look at how all of this is relevant to the theory of a multiverse. The theory of a multiverse is that there are multiple universes, each with a different random set of characteristics. Speaking in terms of probability theory, each of these universes may be considered a separate trial. The series of trials in this case is the entire ensemble of universes, or multiverse. The theory has been introduced to try to explain why a universe such as ours (which looks like an incredibly improbable long shot) might exist.
General Term  Meaning in This Case  Normal Thinking  Multiverse Theory 
Number of trials

Number of universes

1 (?)

Very many

Number of successful trials  Number of habitable universes with life  1 (?)  Many 
trial success probability  The chance of any one random universe being habitable  Incredibly small number such as 1 in a billion trillion quadrillion  Incredibly small number such as 1 in a billion trillion quadrillion 
The last line in this table indicates why the multiverse theory (at least in its simple form) is a total bust, dud, and failure from an explanatory standpoint. Because of Principle 4 described above, the multiverse theory leaves us with a trial success probability that is not any higher than the incredibly low number we started out with. If the chance of our universe randomly being habitable is something like 1 in a billion trillion quadrillion before we think about a multiverse, then we have exactly the same incredibly small number after we adopt the multiverse theory.
In other words, the multiverse theory does nothing to make the “miracle of habitability” of our universe seem any less miraculous. It's the same type of “ten consecutive royal flushes in spades” type of unlikelihood, even after we adopt a multiverse. The longest of long shots of our universe being habitable is just as long a long shot even after we assume a multiverse.
To help clarify how great a weakness this is in the multiverse theory, let us look at a wildly speculative theory that would not suffer from such a defect (although it would suffer from a different problem). I can call this fanciful theory the viral multiverse theory. The theory can be fancifully expounded as follows:
Once upon a time there were a vast number of uncreated universes, each with a different random set of characteristics. Purely by chance, one of these universes was habitable for intelligent life. Then, through some weird strange process, that universe infected another universe with its favorable physics, making that universe habitable. Then the infected universe itself infected another universe with its favorable physics. Habitability thereby spread like a virus throughout all of the ensemble of universes. When this long infection period ended, every single universe in the vast ensemble of universes was habitable.
Now let's expand the previous table to include this fanciful theory.
Normal Thinking  Multiverse Theory (Regular)  Viral Multiverse Theory  
Number of universes

1 (?)

Very many

Very many

Number of habitable universes with life  1 (?)  Many  Very many (all trials successful) 
The chance of any one random universe being habitable (trial success probability)  Incredibly small number such as 1 in a billion trillion quadrillion  Incredibly small number such as 1 in a billion trillion quadrillion  100% (all universes end up habitable) 
When we compare the regular multiverse theory to this viral multiverse theory, it helps clarify what an explanatory failure the regular multiverse theory is, and how lame the theory is from the standpoint of explaining anything. Fanciful and ridiculous as it may be, the viral multiverse theory leaves us with an explanation for why our universe is habitable, something which the regular multiverse theory completely fails to do (as the trial success probability does not change after we believe in a multiverse).
The believer in the viral multiverse theory can say, “Why of course our universe is habitable – all of the universes in the vast ensemble are habitable.” The believer in the regular multiverse theory can say no such thing, and in that theory our universe being habitable is still the longest of long shots.
So why can't we just adopt a theory like this viral multiverse theory? It's because the idea of one universe infecting another universe with favorable physics is absurd. No one has the slightest idea of how one universe could infect another universe, causing the second universe to have favorable physics like the physics of the first universe. Imagining such a thing seems like an invalid case of applying an idea from biology into the realm of physics where it has no business existing. No one even has any workable idea of how there could be any contact whatsoever between one universe and another universe. Plus there are cosmological reasons for thinking that our universe has had favorable physics from its earliest beginning. Numerous scientists have noted that if the Big Bang hadn't been just right, the universe would either have expanded too fast for galaxies to form, or the universe would have collapsed in on itself or collapsed into black holes. So we can't really imagine that our universe started out as “any old universe” and then got its favorable physics long after it originated.
The multiverse enthusiast therefore has a Hobson's choice, a choice between equally unattractive alternatives. He can choose a regular multiverse theory which has no explanatory power because it does not change the incredibly low chance of our universe being habitable, or he can choose some weird variation of the multiverse theory that might explain why our universe is habitable, but at the price of requiring you to believe in some crazy, laughable idea such as a multiversewide viral crosspollination of universe physics (or some other equally ludicrous and byzantine piece of conceptual baggage).
So where does this leave the person who wishes to explain away the astonishing finetuning that our universe seems to have? That person is left in an uncomfortable place, stuck with a deep mystery (“the miracle of habitability”) he can't explain away with a plausible physical theory that gets the job done with explanatory rigor.
Are the cosmic dice loaded?