In this case there is no smoking gun either way and a natural origin is much harder to trace than lab origin. That lack of absolute certainty, which science is often comfortable working with, leaves room for alternative theories to circulate and attach to those who like the sound of them for various reasons. It is a common human flaw that being aware that an event could hypothetically occur can be mistaken for proof that event actually occurred.
Why do you think that? Natural origin should leave closely-related viruses in nature, and animal sampling should find that. Such evidence was found for both the original SARS (palm civets) and MERS (camels) within about a year. For SARS-CoV-2, after about the same time and a much more intensive effort, we're still waiting.
Lab-accident origin is easy to trace only if the people working in the lab disclose everything they were working with. This requires both perfect honesty and perfect knowledge. The WIV did lots of work sampling viruses from nature, in remote bat caves that very few other humans would ever enter--around Yunnan to be clear, about 900 miles from Wuhan--and they could have leaked a novel virus before they even had a chance to sequence it. Of course that's still more likely to imply someone's lying, thus the accusations of "conspiracy theory"; but by that standard every human deception in history is conspiracy theory, including every intelligence operation, every Ponzi scheme, every cheating spouse, etc.
I'd put lab origin and natural origin around equal probability myself, and I'd consider anything between 10/90 and 90/10 reasonable. What would you estimate?