A new page that is demonstrably better than what ranks above it will often fail to displace it for months. The reason is that ranking uses inputs which take time to accumulate.
Links accumulate rather than arrive
A page that has existed for years has been referenced by whoever needed to reference that subject during those years.
A new page has no such history, and the references it will eventually attract have not happened yet.
Because links remain one of the stronger ranking inputs, the incumbent holds an advantage that quality alone does not overcome quickly.
The advantage also compounds, since a page that ranks well is the one people find when they look for something to reference, which earns it further links.
Behavioural history is an input a new page lacks
Search engines observe what people do after clicking a result, and a long-standing page has accumulated a large record of that behaviour.
A new page has no record, so the engine must estimate its likely performance rather than measure it.
Estimates are conservative, which means new pages typically enter below where they will eventually settle and rise as data accumulates.
Ranking systems are biased toward stability
Reordering results has a cost, since a change that turns out to be wrong degrades the experience for everyone searching that query.
Systems therefore require a degree of confidence before promoting an unfamiliar page above an established one.
This produces the familiar pattern where a page appears briefly at a high position, falls back, and then climbs gradually over subsequent weeks.
Updating beats replacing
Because the advantage attaches to the address rather than to the words, revising an existing page inherits everything it has accumulated.
Publishing a new page on the same subject splits the accumulated signals between two addresses and forces the new one to start from nothing.
Substantial revision of an old page, keeping the address, is consistently the faster route to a better position than writing a fresh one.
Where age stops protecting
Age is not a ranking factor in itself, it is a proxy for accumulation, and accumulation decays when a page stops being referenced or stops satisfying searchers.
Pages that have become factually outdated lose position quickly once the intent behind their query shifts toward current information.
In fast-moving subjects the accumulated advantage is small, which is why recency-sensitive queries turn over far more readily than reference ones.