KorZen's Speedrun Strats Deep-Dive!

In these videos, I explain all of the small strategies and tactics I use when speedrunning the Any% category. The first four videos explain the strategies used in my 10:48 world record. The 5th video is an update on every level that changed for my 9:51 world record, and same for the 6th video but for the 9:29 instead.
1-19: https://www.youtube.com/watch?v=Pu2_WoZ1mQA
20-29: https://www.youtube.com/watch?v=zGEN_tlTJPc
30-41: https://youtu.be/08jczbWOaAw
42-52: https://www.youtube.com/watch?v=wM_bxWkH7vU
Updates for the 9:51 record: https://www.youtube.com/watch?v=mDOkNhe_tEU
MORE updates for the 9:29 record: https://www.youtube.com/watch?v=ld_VUAvioG4

Reset exploit - the discovery, explanation, applications, and effects

On November 3rd, 2018, I got two runs of BFDIA 5b with the same exact in-game time, yet it took a different amount of time in real time. So, how does this happen? This is an exploit of the in-game timer that I found, which I am calling “reset exploit”. I discovered this while exploring the game and practicing speedruns of it sometime in mid-September 2018.

On the level select screen, at the top right, there is a timer. This in-game timer has been used for most 5b speedruns so far. I found sometimes I would spend a lot of time in the level, yet only a few seconds would be added to the in-game timer. If you reset a level, the in-game time does not include all of the time before your reset in the level, so it does not add it to the in-game time.

It will add time if you either finish the level or press “menu” to go back to the level selection screen. Also, the in-game timer still runs if you switch to another tab or program, and moving obstacles will still move.

This is a game-changer for both full-game runs and individual level runs. This is the reason why if you play the game with a separate real-time timer, you will take longer in real time than the in-game time if you reset at all during the game. With individual levels, you can practice and reset as much as you want if you mess up, then only your last run will count for the in-game timer. It doesn’t matter if you spend 10 seconds, 1 minute, or even an hour on the level, as long as you reset, only the last run will count. In fact, this is why you’ll see me reset as much as I want during individual level runs.

Theoretically, in full-game runs, you can play levels while resetting if you play sub-optimally, then only finish with good time added to the in-game time. However as a side effect, this does however makes it less of a “speed”run in real time. In fact, my first world record of 21:53.1 took about 2 hours since I kept on resetting, trying to beat individual levels as fast as I could. Reset exploit lead to the situation where I was frustrated throughout the run about being behind my RTA personal best, yet because of the time saved from resetting, I ended up getting a new in-game time world record because when taking out the time before the resets, I was playing faster than every other time I had played before. So even though in real time I was over a minute slower than my personal best, I still set a new world record, which made me very surprised.

So if we keep trying to lower the in-game time world record, eventually the best way to play it would essentially be an individual level marathon. It could take several hours, even more than a day to get the perfect times on levels. I know this all seems a little exaggerated and outlandish, but this could be the future of speedrunning BFDIA 5b in order to bring the world record lower. Even if resetting for hours in order to get a better time sounds silly, that would be the best way to get a lower in-game time, and getting the shortest time is the goal of speedrunning.

What’s the solution? I suggest making a sub-category for existing full-game run categories where real-time attack (RTA) and in-game time (IGT) are separate. Individual level runs are often short enough it’s not worth timing them externally, so i think the existing method of looking at how much your in-game time went up should work now, even if you have to do some math each time.

In-game time isn’t necessarily wrong, both RTA and IGT have their pros and cons. IGT works best if you can’t record your runs since it can be recorded in a screenshot, and could possibly help time if your computer is. slow. However, as I mentioned earlier, trying to go for the in-game time world record could be a problem in the future. In-game time also only goes to the tenths of a second instead of milliseconds. Since we can’t see past those decimals, trying to save decimals of a second on short individual level runs could be a problem since you can’t tell if the in-game timer is as low as .x00 seconds or as high as .x99 seconds. But if you’re just trying out and don’t want the hassle of setting up a separate timer or recording software, looking at the in-game time should be good enough.

I want RTA to be the future of full-game runs, since that’s how most speedruns are done now. I don’t want 5b speedrunning to turn into individual level marathons for full-game runs, as that’s kind of the opposite of a “speed”run. Sure the RTA will be slower than the IGT, but I think it’s more practical this way. I’ve already started using an external timer and I like using it and seeing the splits.

I hope you all thought my discovery was interesting. It really does change the  whole game in speedrunning contexts, so I hope we can agree on what we should do with reset exploit. This should make you understand some of my 5b speedruns a bit better.

===

This script was originally written on Nov. 7th, 2018 for a video I was planning to make. I originally submitted it here on Nov. 15th, 2018, but it was removed, so now it's being resubmitted on Nov. 17th, 2018. Since the original posting, the leaderboards have been split into IGT, RTA, and TAS on Nov. 17th, 2018.

Nome	Autore
KorZen's Speedrun Strats Deep-Dive!	KorZen 3 years ago
Reset exploit - the discovery, explanation, applications, and effects	MeesterTweester 6 years ago