Uploaded April 2023 | Updated September 2026, 3 weeks ago

---

[See watch?v=pmw7XRE5E1w for update] quickly recorded these three instances, and put them through editing in thirty minutes or so. Played on JPN.

This video saw birth thanks to the retroachievements leaderboards of this game. To make things simple, after following procedures and downloading programs, you're able to emulate a game, play on it, all the while its data is being read by an external program, a bit like x-rays.
With a lot of craftiness, it's possible to create a system that based on how the data is running like, yields you achievements based on your actions, or gives you access to time-attack, purely based on how the data is running.

And for this game, among other things, there happens to be a boss time-attack implementation for every single boss, including al individual Nahatomb phases. You can play on it and compete for the lowest time in the leaderboard. As follows : youtube.com/watch?v=jg1NDMNw028 (the timing for these bosses are from when the health bar appears and the final hit is administered)

Since forever - there had to be one top level DtP representative, and I was busy -, amoser was the king of the hill, with a fair surplus of time behind everyone else, for most ILs. And as you could imagine, some of his times would be completely off-limits to people unless they knew about the top level speedrunning strategies.

But today, this is different. Because since the beginning of april of this year, two of his times have fallen.
And the faster times... performed by a unknown figure by the username DaniWoX... appear to make absolutely zero sense.

The first time is set on Baladium, with an improvement of roughly 15 seconds over amoser's time 28.93, with a 14.56 : retroachievements.org/leaderboardinfo.php?i=9348
To save seconds of time over complex fights would be one thing. To save 15 seconds with Baladium, an incredibly rigid and locked-in-place fight, whose fastest and only strategy that speedrunners have since the dawn of time for that matter, a child could quite literally figure it out, seems not of this world.
If you ask me, the only way to imagine this time-save as possible, and by a second, would be if you do the following two things :
- Loading the fight by letting the cutscene finish, a very small improvement allowing us to hit Baladium faster (see youtu.be/PkuwvC5XVQg?t=33)
- Bringing Baladium down with only two throws... ?????????????????????????????????????
Don't forget the part where you have to make both throws without failing.


The second time, Nahatomb first phase, 9 seconds over amoser's 49.61 with a 40.31 :
retroachievements.org/leaderboardinfo.php?i=18751
Yet another completely stiff and cycle based fight. Unlike Baladium, its fastest strategy wasn't as readily known and as obvious, but speedrunners would still have quickly figured it out anyways.
But just like Baladium, it remained the strat every top players were more or less relying upon.

Suffice to say, these times don't sound reasonable. What's more, this player's other times are unremarkable in comparison. Ask me what's the odds of coming up with not just one, but *two* brand new strategies on some of the most immovable bosses, possibly insanely difficult strategies, but not having the slightest clue at how to do the other fights.
Can a three minute time owner of Balue's Tower really do this ?

It may be blatant cheating, or more reasonably, a bug with the game data reading that causes the IL to end prematurely, like how it had been in the past of this leaderboard. Unless we contact this player, or find these "strats" on our own, we may never know for sure if it's real or fake.



...But with this said and done, I got thinking.
Staying on imaginary bounds, there was perhaps a way to save this much time on Nahatomb. The only way we know : by skipping an attack. Specifically, my mind fixated on the tongue attack.
I always thought it was completely impossible. Because even with perfect movement and planned positioning of the three remaining death Moos once their armor get broken by Nahatomb's fist shockwave, he would initialize the move well before you can fill the fourth cannon. Or at least, trying to abide by the standard approach of the fight, of filling one cannon at the beginning, then three more using the running death Moos, and then a 50/50 normal death Moo spawn with the last cannon.

And that's when I realized it. What if you simply didn't go by this order ? What if you made it so you force the tongue attack to never show up, with strategic enemy handling ? Could we hope that this is faster ?1
Well, on paper, it was hard to gauge, but surely intriguing. I quickly jumped in the game to try it out (displayed on the left).

Trying this one approach, I encountered a variant of it (right), as it's still following the spirit of not having the tongue move show up.

I do not have much knowledge with these strategies, nor of their consistency factors. But these are unequivocally faster : 6.8 seconds.

To continue...