We’ll never get an official answer, instead I’ll just repeat what I need to set up so I can copy/paste in any one of these threads:
It takes a lot of data to prove a % is off, but there are more ways to be “off” than raw statistics.
A long time ago, TDS had a 50% resurrection chance and players insisted it procced more often than 50%.
Technically it was the wrong trait, but they didn’t have time to make its real third trait. It stayed this way for something like a year, it’s hard to find the time to write one trait.
Devs insisted the players needed to stop spending money on booze and spend that money on gems instead because the RNG was fine.
A player with a very good basis in stats performed a very detailed analysis and concluded:
If you look at proc rate in aggregate, it procced at 50%.
If you inspected it via statistical analysis designed to detect streak biases, the trait was far more likely to trigger 0 or 3 times than an unbiased RNG would suggest.
TL;DR: TDS was proccing more per match than it should and making up for it by “not proccing” more to disguise it. A probability fairy. Or, in layman’s terms, “a sign the RNG has bias”.
The devs suddenly found time to implement the “correct” trait and applied it to TDS, because that’s not suspicious.