This is how the entry "Most moves in a chess game" was presented by Guinness:

The Master game with most moves on record was one of 269 moves, when Ivan Nikolic' drew with Goran Arsovic' (both Yugoslavia, now Serbia and Montenegro) in a Belgrade, Yugoslavia tournament, on 17 February 1989. It took a total of 20 hr 15 min. There is now a 50 move rule so unlikely ever to be beaten. This record is sourced by the World Chess Federation. We do not send out guidelines.
Let us focus on the phrase: “There is now a 50 move rule…” The 50-move rule was in force then as well! Who misled Guinness World Records?
Now let us look at the concluding sentences: “This record is sourced by the World Chess Federation. We do not send out guidelines.” It reads like a disclaimer: “This information was provided by FIDE. We are not responsible for it.” For example, the information about the “Longest chess marathon” is published without any such disclaimer.
What prevented the Yugoslav Chess Federation from double-checking the Nikolic–Arsovic game before submitting it to FIDE? What prevented FIDE – whose president at the time was International Master Florencio Campomanes – from double-checking the game before forwarding it to Guinness World Records? These are no longer rhetorical questions!
A careful study of this topic has led me to spot a mistake in Part 4 of my article. I stated that the 100-move rule applied to three types of endgames. In fact, it applied to eight:
In 1984, this rule was introduced as an experimental innovation (Article 10.9), and disputes immediately arose as to whether it could be applied to “similar endgames.” As a result, in 1988 a consolidated edition of the Laws of Chess was published, in which the 100-move rule was integrated into the main body of the text. Most importantly, the rule explicitly stated that it applied exclusively to the eight listed endgames.
It is also crucial to note that in pawn endgames the rules specify the rank on which the pawn may stand. For example, in the endgame K+Q+P vs. K+Q, the pawn must be one step away from promotion, while in the endgame K+N+N vs. K+P the pawn must not have advanced beyond the so-called Troitsky line.
This means that in February 1989, when the game Nikolic–Arsovic was played, the endgame K+R+B vs. K+R+P was not included in the “list of eight”, and therefore the 50-move rule, not the 100-move rule, was in force!
Thus, I repeat the conclusion reached in Part 4 of my article: “The game Nikolic–Arsovic was deliberately prolonged.” It should have ended after move 161, not after move 269. [Spoiler: even on move 102!]
This game was adjourned many times (at least four), and judging by the tournament table shown in Part 4 of my article, everyone connected with the event was convinced that the game would enter the Guinness Book of Records – not only as the game with the greatest number of moves, but also as the longest game in terms of actual time spent at the board.
With 269 moves and a time control of 2 hours for 40 moves plus 1 hour for every subsequent 20 moves, after the game ended each player still had approximately 4 hours left on the clock: ((280 / 20) × 2 − 20 hours of play) / 2 = 4 hours per player.
This raises an obvious question: why did they not use all the available time? Were they confident that 20 hours and 15 minutes would suffice for the record? But 28 hours would have been far more impressive!
However, this is not the only aspect of the game that raises questions. Why play the endgame K+R+B vs. K+R at all, if it is a theoretical draw? Exactly! It is a draw – but a difficult one, and the weaker side can make mistakes.
But let us proceed step by step.