One method I could think of is playing the game on Flashpoint. To my knowledge, you can't modify the game there and the item eating glitch works there.

Huh, I didn't think it was still possible to modify Flash games since it got canned.

From what I remember during my code digging, item drops are not attached to enemies but rather the battle itself. I'll have to look into the 'itemRareDropper' to see if it does anything special.

I did wonder myself where exactly the cutoff was for 100% when I ran that category, so I'd say trying to set some rules for that is warranted. I agree that option 2 is more consistent to re-time and less likely to cause arguments later.

The issue with using an Assassin build for the 100% route is the stat gains from paladin armour. Paladin armour buffs Lightning Piercing and not Physical Piercing, which obviously is better suited for a Gunslinger route firing Electro Bolts. This matters quite a lot on the final battle of Zone 4 as the Assassin build doesn't quite have the damage needed to beat it while the Gunslinger build does.

On the topic of the timing of attacks, I did notice that magic attacks have faster animations compared to physical ones. Probably the biggest example of this are with the ZPCI and most enemies with guns, since their SWAT Fire attack has maybe the fastest attack animation in the game.

It's been over a year since my WR and I have yet to come up with a strategy that doesn't overly rely on RNG, so I haven't played the game in a while. I feel like at this point, a new glitch needs to be found in order to improve all routes.

I was thinking about how I could make the any% run more consistent and came across an idea.

Currently, we do training fights in Zone 3 in order to grind for equipment and levels. In the best case scenario, assassins would fight Shaman of Blades for Tribal Blades and Gunslingers would fight Shaman of Life for Mender's Axes. However, this is very inconsistent as there is only a 25% chance of fighting the desired enemy each time + a 25% chance they drop the weapon, not to mention time-consuming as these enemies are difficult to kill right after Zone 2.

The idea for the assassin route is to do most of your training in the beginning of Zone 2. The training enemies all have a 5% chance of dropping Moon Shattered Horrors and are relatively easy to kill without taking too much time. If you do 5 training fights, you have a 22% chance of getting at least one MSH which will make the rest of Zone 2 much easier to finish + eliminate training in Zone 3.

The main drawback though is that the items drops in general for these fights are terrible. There's only a 30% chance that each item that's not MSH will drop, so there are plenty of fights where you get nothing from it.

I did some testing and at one point I got a 14:04 (with no recording though).

So I tried some preliminary testing using Master Strike in any%.

It's not worth. Definitely not worth.

4/5 Master Strike can't out-damage 3/5 Smash even when I spent all of Zone 3 dumping points into speed through level up and items. To say I'm disappointed is but the tip of the iceberg.

Also, dylaneka is the only active mod we have left. They're usually pretty quick to approve runs too.

I agree that the most fun aspect of speedrunning this game is finding what route is most optimal. Probably the biggest aspect that has to be addressed in any% is what abilities you use and when do you gain/upgrade them. Usually it's only after you decide on your abilities that you start to worry about the class and items that you have to eat. However, I feel that abilities are the most restrictive part of Sonny.

For any% you need to quickly establish your offensive core of abilities and stick with it for the entire game. Respecing causes all the stat boosts from item eating to go away, so you can't use fully-upgraded Smash in the beginning and then respec to a fully-upgraded Master Strike/Coup de Grace + Shatter Bolt Build later (if you even have enough ability points for that). The only two sets of abilities that can form an effective offensive core early that I can see are Smash for physical builds and Magic Bolt + Electro Bolt for magic builds. No one has done a comprehensive test for how the other abilities perform, but glancing at them they're all rather pathetic. They either suffer from being too costly to obtain like Master Strike or being awful to upgrade. Most abilities don't scale with Sonny's stats, but rather by some constant number which makes upgrading them a waste of ability points. For example, Blood Strike:

• Tier 1: Damages target with 110% of caster's Strength +60
• Tier 2: Damages target with 110% of caster's Strength +75
• Tier 3: Damages target with 110% of caster's Strength +90

There IS an idea I'm thinking of right now for a re-route, but it requires testing that I don't have time for. Plus it may not even be that effective, so whoop :)

So I whipped this video together in an hour. Hopefully it can serve as a start to speedrunning the game! There's also the written guides that you can read, and you're also free to ask us questions~

Oh I totally forgot, as far as I'm aware it's totally okay to use Flashpoint to play the game and submit runs with. With flash on browsers down it's one of the only ways to play the game right now.

Yooooo a new runner! Welcome~

This game can be quite deep when it comes to speedrunning despite it's short length and comparatively simpler battle system to other games. The fact that it's a point and click game demands excellent menuing and making minimal mistakes, and a lot of thought goes into the decisions you make in each fight. Overall, it's a charming game to speedrun whose main downfall is the RNG involved with item drops.

Tbh, My WR run is not a good way to learn how to run any% using the assassin class. The sheer luck in it completely eliminates a lot of the strategic decision making that goes into many fights. I've been wanting to make a video guide or write one here because of it.

Let me know if you have any questions about the run. I would love to help!

Hmm...I agree that using the values of the attributes instead of counting the number of items might be better. In that case, we could maybe count it like this:

• The highest value that each stat reaches in the run
• Abilities acquired
• Maybe allies

By using the value of each stat, we eliminate all of the other categories I suggested before. I feel that Veradux could be a big part of what makes a run possible. His healing and focus regen could be the make-or-break of a fight, so counting him is probably best.

As I continue to stress over university assignments, I sometimes find myself wondering what a Sonny low% speedrun would look like, or if it would even be fun.

As far as what counts as a 'thing', I've come up with the following list:

• Levels gained (Including the percentage too? If it comes to it...)
• Attribute points used
• Abilities acquired
• Items equipped/eaten via glitch

There's also the use of teammates, where you are forced to use Veradux and Galiant in their first battles (but Veradux can be removed later). Then there's the initial equipment that Sonny has, which I guess can be removed after the first battle. We would have to discuss whether the fact that we were forced to use them once would count them in the overall total of 'things' used.

I hope this sounds interesting! :)

Just to add onto the consequences of Flash being terminated on Windows, it's put a halt on any progress I can make into looking at the games code. To be precise, I can still LOOK at the code, but it's now impossible to check how any changes I make affect the game (not that I was actively making any progress, but hey lol)

So as promised, I've been looking into Sonny's code for the past few days and have some things to report (and to not report).

The first is as to how the game determines what items drop after a battle. As I feared, it uses Math.random() as the powerhouse behind its 'randomness'. The reason this is a problem is because you can't seed Math.random(). Seeding basically means that if you input the same 'seed' into the random number generator, it would output the same results. So because of that, at the moment I have no idea how to consistently get the drops that we desire (MSH plz).

The second thing, and this didn't really have to with speedrunning but my inner code geek came out, was that I found out how the game creates items and enemies/allies. Nothing much here that's useful for routing, but it was fun to mess around with item stats and flavour text. Here's an example of what I did: https://twitter.com/AzureStarPlayer/status/1339138161872576512?s=20

And finally, the formulas for how the game calculates damage. The good news is that I'm pretty sure I found them. The bad news is that I'm going to have to do some debugging in order to find out how it works. The issue is mainly due to me not really understanding the structure of actionscript. There are variables in the formulas that are declared somewhere that I can't see, and therefore I don't know what they are or what possible values they should have. My plan right now is to have the program spit out the values for those variables as I play the game, and hopefully from that I can deduce what they're meant to represent.

So that's what I've got so far. Nothing too impressive, but hopefully I'll have something in the coming weeks.

Hmm....should I help with this? Technically I would like my WR to remain standing for as long as possible XD

Just kidding, I was going to focus on this anyways so I'll get down to work. Yesterday I was poking around the games code, and while I'm completely new to actionscript, I am familiar with Java so as long as I can find the right scripts then I can get the formulas that calculate damage.

I've also been trying to find out how the game determines what items drop, but a preliminary search suggests that it's based of a Math.random number which is not good.

Keep in mind though that this is my first attempt at looking through code to determine glitches/secrets/exploits so don't expect anything revolutionary, or for my findings to be entirely correct.