Recently, on-chain data always seems to "pause" for a moment, it's not your network connection... Many times it's the indexer/subgraph catching up with blocks. When a new block arrives, it has to fetch logs, run a mapping, and then write to the database, which can cause a delay of a few seconds to tens of seconds. No matter how much you refresh the front end, you can only watch in frustration. What's even more annoying is RPC rate limiting, especially during airdrop season when everyone is grinding tasks like clocking in at work. The task platforms also fight against bots, with a bunch of scripts, and public RPCs get overwhelmed. Your side just shows request timeouts, and balances/positions appear and disappear intermittently. Anyway, when I see "data jumps," I don't blame the smart contract first. I switch RPCs, close a few pages, or wait for a couple of blocks to confirm, to avoid scaring myself.