IIN I wonder why atheists are so confident?

Empirical sciences work by accumulating and, most of all, replacing tentative, imperfect, probabilistic evidence with new evidence obtained by tests that undermine at some level previous evidence. Unlike other believer stated earlier here in this thread, no, there is no faith involved, but all of it is – and its credibility (the highest there is, on this I agree with you) is based on this fact – inherently provisional, imperfect, probabilistic and incomplete. Even the "laws" of Physics are just imperfect models, super accurate and predictive as they may be. Newtonian laws seemed just as "perfect" as the current laws seem right now until we started to develop the technology necessary to make tests in systems of greater scope, when it became obvious that they were pretty unreliable at that level. Even so, at the level they worked before with very little error, they still work, even though their wrongness now is so much clearer. Of course, current laws are more accurate, but they are still imperfect, they don't explain everything we observe, and odds are the laws we arrive at in the future won't either. There are physical "laws" we don't know yet, some we may never know, and the laws we "know" now may (most likely do) have better formulations and explanatory foundations than the ones right now available to us. Does that mean these results and explanations are bad, unreliable or equivalent to theological statements? Nope, they are still more reliable and predictable, obviously, as I stated before. But they still don't prove anything, they just accumulate tentative (sometimes, like in this case, very good) evidence, and of the many things they accumulate evidence about, well, God is not one of them. We can, yes, accumulate evidence about things that in Theology are related to narratives about God, and these things can refute some theological statements, but both religion and Theism survive all of this, because religion is dynamic and change many of its beliefs with time, while some of them, like the existence of God, don't need to change, since they are untestable and thus out of reach for the tools of science. God is not equivalent to a spaghetti monster either, since the sociohistorical background, philosophical foundations and evidence are still much stronger and its anthropological status completely different. However, the evidence (or lack thereof), indirect though it may be, for the existence of the Christian God that is the strongest in theological narrative happened in the far past, 2000 years ago, and the other minor evidential reports were also either far in the past, or unrecorded by scientific standards, or not within the minimum requirements of the empirical sciences. In other words, using God as an explanation goes against Occam's razor, at one level, since it does not stem directly from supported (that is, tested) evidence, and at another level is in itself untestable and in this sense uninteresting for science, while “God’s will” variables cannot be observed in data in any significant way. Thus, the rational reason for not believing in God is not that science proves He does not exist, nor that science is all encompassing at any given point in its knowledge (both claims are unarguably false), but that God is not falsifiable, that is, impossible to test, and thus uninteresting from a scientific perspective, and its suggested existence based on evidence itself unfalsifiable, while He cannot be directly observed in trends that cannot be explained by operationalization, which means that, from a scientific point of view, there is no reason at all to rely on God as an explanation, nor to actively believe he exists. Since science is itself linked to probability, most of the models of decision-making based on probabilities assumed from the given scientific data are also not supportive of adopting a belief in God, based on the same arguments above. Many actually argue that Occam’s Razor can basically be understood as a less formal but more theoretically grounded form of this very same style of modelling probability. Does not mean this actually says anything directly about God or comes close to refuting His existence, just that, on an evidence-based model, the reasons for believing in God with the best knowledge of data are weak.

By the way, I may be a believer, but the best atheist philosophers are going to give you basically the same answer, but probably with much more and better detail into the philosophical considerations involved, and probably with less mistakes regarding the philosophy. As a scientist, I know only the basics on that front, and this is basically a philosophical, not scientific, problem. By the way, Dawkins is a very good Evolutionary Biologist, but a bad philosopher. William Lane Craig, the believer who crushed him in a debate, is a better philosopher (good enough to know at least some of the major arguments relevant to the problem, and Dawkins didn’t; that’s why he was crushed besides being mostly right), but IMO still bad (again IMO, he arrives at the wrong conclusions and is one of the main models in a school of theology that goes against many of the arguments above), besides being basically a conman (since he just cherry-picks popular amateurs in philosophy like Dawkins in order to crush them on video and make himself look good, when he is not actually engaging the best in the field at all). If you want an easy to find, generally good resource on problems like this, I would check the Stanford Encyclopedia of Philosophy, then look for relevant authors representative of the major arguments mentioned there and in other credible sources.

Now, why one believes is another question not touched above, but most answers in Christianity will turn to faith at some point, at some level and according to a theological conception of the term. Like I mentioned, it’s not the ideal option from a purely probabilistic point of view, but it may be in accordance with other reasons to believe (and, after all, almost no one follows the ideal probabilistic option in most of one’s life). Few atheists even know the best probabilistic reasons to be an atheist, and everyone is full of at some level irrational thought processes or solutions, usually with much more trivial reasons for their adoption than those taken into account for the faith in God.