Concrete or gravel is the wrong question. The question is what happens to the water.

The field evidence points at the soil line, not the concrete

This is the most useful thing on the page and it comes from somebody describing posts they had just dug out. Their sixteen year old posts had failed where the soil was in contact with the wood above the concrete, while the wood inside the concrete was still very sound. And in the sections where the old concrete bases stood taller and were not topped with soil, the posts were sound too. Read that carefully, because it inverts the usual story. The rot was not caused by encasement, it was caused by a permanently damp soil-to-wood contact line a few inches above the footing. Which means the practice to change is not necessarily using concrete, it is burying the concrete under dirt so that such a line exists at all.

Crown the footing above grade, and keep soil off the wood

Following directly from the above, this is the cheapest and highest-value step in setting a post. Bring the concrete up above the surrounding grade and slope its top away from the post so water runs off rather than sitting at the joint. Guidance confirms that sloping the top of the concrete away from the post sheds water and helps prevent decay. Then keep it that way, because grade creeps: mulch gets topped up, beds get built, soil washes against the fence, and five years later there is dirt piled against wood that used to be clear. Whatever you decide about what goes in the hole, this is the part to get right and the part to check periodically afterward.

Leave the bottom of the post open

A specific instruction from a fence contractor and the direct answer to the cup-and-saucer problem: the bottom of the post should sit on either the ground or gravel rather than being encased in concrete, and when you concrete a post you leave the bottom open so the post and the surrounding area can drain. That single change turns a closed vessel into something water can pass through. A few inches of clean gravel under the post achieves it, and it costs almost nothing on top of a job you were already doing. It also means that if you ever do have to replace that post, there is a drainage path rather than a solid plug.

The gravel has to be angular, and river rock is the wrong product

Not all gravel does this job. What you want is crushed, angular rock, commonly sold as three quarter inch clean or crushed drainage rock, and the reason is mechanical: the sharp interlocking edges are what let it both drain and hold. Smooth rounded river rock drains but rolls, so it does not lock together and does not support a post against lateral load. This is an easy substitution to make at the yard because the bags look similar and one is often cheaper, and it is the difference between a gravel-set post that stays plumb and one that leans in the first year. A practitioner method worth copying: a few inches of clean gravel in the bottom of the hole, then the post, then about six inches more clean gravel around it, tamped.

Gate posts and corners are a different decision from line posts

This is where the argument stops being an argument. A gate post takes a repetitive swinging load and a corner post takes constant tension, and both need rigidity that gravel does not provide as well, particularly in soft or loose soil. Concrete is the right answer there and practically nobody disputes it. Which means the sensible plan is not one method for the whole fence but a method per post: concrete for gates and corners, and a genuine choice for the line posts in between based on your soil and climate. Deciding this per post rather than per project is what a professional does and what a homeowner following a single blog post usually does not.

Your soil decides more than your preference does

Guidance is consistent that this is soil-dependent rather than universal: clay soils and taller fences benefit from concrete anchors, while well-drained sandy or rocky soils may work with gravel alone. Clay is the specific problem case, because it swells shut and traps water directly against the wood, so the drainage argument matters most in exactly the soil where gravel drains least well into the surrounding ground. Frost cuts both ways too, with gravel reducing frost heave while concrete itself can be affected by it and can crack or shift over time. Dig one test hole before committing to a method, fill it with water, and see how long it takes to disappear. That single observation is worth more than any general rule.

Local conditions can invert the whole question

Worth including because it is a useful check on treating any of this as universal. An owner in northern Alberta describes ground frozen from September to June, bedrock anywhere from a few inches to three feet down, and seventy year old posts that rot from the top down rather than at the base. In that climate the ground-line moisture problem this page is largely about barely operates, and the constraints are frost, rock and a four week window when the ground is diggable. If your conditions are genuinely unusual, local practice beats general advice, and the way to find local practice is to ask somebody who has dug posts out of your soil rather than somebody who has put them in.

The wood you choose changes the stakes

The setting method interacts with the material, and the range is wide. Fence Armor puts an untreated post in concrete at only five to ten years, and a treated post with proper drainage and protection at twenty-five years or more. One contractor’s figures give cedar in gravel twenty or more years against twelve to fifteen in concrete, with cedar outperforming pine in either. So the same hole treats different posts very differently, and the drainage decision matters most for the least rot-resistant wood. If you are using untreated or borderline material, the drainage detailing is not optional refinement, it is most of what determines how long the fence lasts.