The arguments you'll actually hear — and a short answer to each

“Everything that begins has a cause; the universe began; so it has a cause”

Reply: The first premise is drawn from things inside the universe, where causes come before effects in time; it's not established for the beginning of time itself. The Big Bang model describes expansion from a hot, dense state, not an absolute beginning out of nothing. The Borde–Guth–Vilenkin theorem is often cited here, but it only shows that a certain kind of inflating spacetime can't be extended infinitely into the past. And even if it worked, the argument gets you to “a cause”, not to a particular god, let alone the God of any one Bible.

Where it has a point: “Why is there something rather than nothing?” is a real open question, and “we don't know” is the honest answer.

“The constants of physics are fine-tuned for life”

Reply: To say a value is improbable you need a range it could have taken and a probability distribution over that range, and nobody has either: we have one universe to look at. We also only ever observe a universe that permits observers, which is a selection effect. And most of the universe is hostile to life, which fits poorly with a design aimed at life. “God did it” also doesn't predict these particular values; it would fit any values.

Where it has a point: Why the constants have the values they do is an open problem in physics. Multiverse proposals are speculative, and you shouldn't lean on them as an answer.

“Science can't explain X, so God must have done it”

Reply: The gapes have a track record. Lightning, disease, the motions of the planets, and the diversity of life were all once credited directly to gods and now have natural explanations. “We don't know yet” is a complete and honest answer; it isn't evidence for any particular answer. Ask what the positive evidence for the proposed explanation is, besides the absence of another one.

Where it has a point: Some design arguments (e.g. “irreducible complexity”) claim positive evidence, not just a gap. Take those on their specifics rather than waving them off.

“Evolution is only a theory”

Reply: In science, a theory is a well-tested explanation that ties together many facts and makes predictions, not a guess. Gravity and germ theory are theories too. The everyday meaning of “theory” (a hunch) is a different word that happens to be spelled the same.

Where it has a point: Nothing in science is proven in the way a mathematical theorem is. Scientific conclusions are held because the evidence is overwhelming and they can be revised if it changes.

“You weren't there, so you can't know what happened”

Reply: We infer past events from evidence they left behind, which is how courts convict from fingerprints and how investigators work out what caused a crash. Applied consistently, the objection would also rule out the Bible's own history, since nobody alive today was there for that either. Independent dating methods (tree rings, ice cores, radiometric clocks) agreeing with each other is exactly what lets the inference stand without anyone being there.

“Without God, there's no basis for morality”

Reply: Plato's Euthyphro question still applies: is something good because God commands it, or does God command it because it's good? If the first, goodness is arbitrary; if the second, goodness doesn't depend on God. Believers and non-believers alike also ground morality in wellbeing, empathy and cooperation, and moral views change over time, as slavery shows, often against what scripture was used to defend.

Where it has a point: What morality ultimately is remains debated among philosophers on every side. Say that plainly rather than claiming it's settled.

“Believing is the safe bet: little to lose, everything to gain”

Reply: Which god? Thousands of mutually exclusive religions each promise rewards and threaten penalties, and the wager doesn't favor one. You also can't simply choose to believe, and a god who values honesty would hardly reward pretending. Choosing the wrong religion costs something too.

Where it has a point: It's a famous early use of decision theory, and Pascal was well aware of some of these objections. It's an argument about what to bet on, not evidence that any god exists.

“I've felt God. I know he's real.”

Reply: Take the person seriously: the experience was real to them. But people in every religion report powerful experiences that point to incompatible gods, so the experience can't settle which one is true. Brains can produce such experiences without any external cause, and being sure about it is not the same as being right about it.

Where it has a point: Don't tell someone their experience didn't happen. The disagreement is about what best explains it.

“Atheism is just another religion”

Reply: Atheism answers one question: do you believe a god exists? It has no scripture, rituals or creed, and atheists agree on little else. Not collecting stamps isn't a hobby.

Where it has a point: Someone who claims “no gods exist” is making a positive claim and takes on a burden of proof for it. Many atheists make the narrower claim that they've seen no good evidence, and that is the position easiest to defend.

“You can't prove a negative”

You can for specific claims (there is no elephant in this room), and the person asserting something carries the job of supporting it. Ask “what's your evidence?” before getting pulled into disproving it.

When the goalposts move

If a point is answered and the claim quietly changes (“well, I meant…”), name it politely: “We started with X, and we've now moved to Y. Can we finish X first?” This keeps a live debate readable for the people watching.

“What would change your mind?”

Ask it of the other person, and answer it yourself. If nothing could change their mind, the conversation isn't about evidence, and it's worth saying so kindly. Many people watching are deciding for themselves, so plain, patient answers matter more than winning.