These days I've been seeing a bunch of screenshots titled "On-chain data proves xxx," and I can't help but chuckle... It's not that on-chain data isn't accurate, but what you see as "on-chain" is often processed by others: nodes are syncing slowly, RPC queues are causing delays, indexers are still catching up on blocks, so that transfer you just saw happen might have actually occurred several minutes ago.

It doesn't matter much in the medium term, but during the hottest emotional moments, those few minutes are enough to lead people astray.

And now everyone is tying ETF fund flows, US stock risk appetite, and crypto price movements together, interpreting them as if they are directly linked. It sounds logical, but when you get to the on-chain side, with delays and selective data display layered on top, the narrative becomes even more prone to overheating.
I’ve gotten into the habit of stepping back when I see things getting too heated—making a cup of tea first, then waiting for the delays to catch up before seeing if it’s the same story.
The word "delay" has really become more and more common when looking at on-chain data.