That's literally just the nature of choosing a random quote. Think about this: say you had 1000 quotes, and you pick a new one (with replacement). After you pick the first quote, there are only 999 more to pick from such that there is no collision, and then 998, 997, 996 and so on. The further you go on, the higher the chances of one of those repetitions, so after choosing 50 unique quotes, there would be a 50/1000 chance that one of those quotes would have already been chosen, which is the same as 1/20. This chance becomes higher and higher, and at the same time your are doing more and more trials.
This issue is not unique to this website, and there was a similar issue to this with iTunes users thinking the shuffling algorithm was broken, when they just didn't expect randomness to behave the way it did: shorturl.at/qEMS3
