You don't know that they had chat restrictions. Perhaps they realize that it takes longer for their actions to catch up to them if they don't say anything in chat about it, because there's no direct evidence, then. But you can still report them for this afterwards. And in cases like this especially, be sure you include a description of what specifically they were doing.
I've posted this quite a few times now, but this was a response I had to a question about why Riot only cares about chat, and I feel it applies here as well:
It's because of the ease of detection/confirmation. Seriously.
They can have automated systems in place that are capable of catching certain toxicity in the chat. And then a case can very easily be made out of player reports for verbal abuse/harassment which can be manually reviewed. These cases just show chat logs and usually generic post-game info (scores, builds, etc.). It's pretty easy to read through several sample chat logs and see that a player is routinely toxic in chat, and hand out a ban for that.
In-game actions are more difficult. Fortunately, they can automatically detect leaves and AFKs (including those who walk in circles in the base to try to evade the system, but don't actually do anything). When players report for other types of in-game behavior, it's a considerably more difficult case for review. From the information that's easily-accessible, you can easily catch the mob boots + 5x zeals Veigar that's 0/15/0 with 2 CS. Cases where a player announces that they're feeding now, and they proceed to die 3 times in the next couple minutes are also pretty straightforward. But more nuanced situations, like an ADC who is 10/3 disagreeing with an action by their allies, and intentionally feeding 3 kills in a row to make a point, though not actually announcing it in chat, is much more difficult to catch. You'd have to basically rewatch entire games to confirm these.
That isn't to say reports for such behavior are meaningless, though. If a player is regularly doing this, and a lot of people report them for it, it will trigger a more in-depth investigation of the player, which will likely lead to eventual punishment. But this relies on consistent negative behavior from that player, as well as consistent reports from others in the game. Preferably with comments providing context to the report so the reviewer knows what to look for.
The tl;dr is basically this: An automated system that catches trolling behavior in game, rather than chat, is much more difficult to make. This just means it takes longer for these actions to catch up to people and result in a ban. Situations like these basically require a hand-review of the case to determine exactly what's going on. And Riot does that, it just takes more time.
