I've reached stage of development where I am constantly ping-ponging back and forth between "This game sucks actually" and "Literally coolest game ever."

Here's the events and things I've been working on that led to these feelings:

Distancing from FNaF:

There's no denying that Mary's Arcade is a FNaF fan game. That would just be silly. However, I feel so detached from modern FNaF that I'm not sure I want the Mary games to be seen as even taking place in the same universe. Because of this, I've scrubbed all FNaF references from Service Pack's scripts and word tables.

If you want to know what cutting FNaF out of my FNaF fan game felt like, it was something like this:

Breaking away from FNaF allowed me more freedoms. After all, do I really need to stick to just five nights? I can have as many nights as I want, perhaps even a second week. That thought led to...

Hard Mode / Week 2

The original Mary's Arcade is broadly considered brutally hard, but believe it or not I thought it was a bit too easy on launch.

I believe this comes from the deterministic rule-driven AI. If you understand how the characters work (or are particularly good at trial-and-error games) each night in the original Mary's quickly becomes a series of consistent mouse movements that always gets you a win. This is especially true for me.

I planned on turning down the difficulty a notch for Service Pack since I want people to complete the game, but I also wanted to keep the difficulty the same or harder for experts. This culminated in a hard mode.

I planned to tie the hard mode's activation to the naming screen like in Undertale, but since changing your name in Mary's is a lot more of a hassle than just starting a new game in Undertale, I decided against that. I was stumped. I thought about having a hard mode toggle on the menu or something else equally as dumb, but after I divorced the game from FNaF I immediately had the idea of just adding another five nights and making that the hard mode.

"So is hard mode just going to be the same game with harder AI?"

No.

Hard mode in Service Pack isn't going to be a "weapons do less dmg / enemies have more hp" mode. I have plans to make Martin - the custom night exclusive from the first game - play a big role in hard mode.

I can already hear the custom night victors from the first game groaning, but don't worry, he's being redesigned from the ground up.

With the game's progression royally screwed up (you know, having 2 weeks?) I couldn't find a good place to put the extras menu. Should you unlock it after night 5 or night 10? The game is in fact over after night 5, week 2 is optional. Eventually I decided to replace 'extras' with an almanac that populates as new characters activate. (After night 1 you unlock Mary, etc.)

Same deal for the Custom Night; you unlock it after night 5, with week 2 characters unlocking as you play.

Fundamentals of fun

Getting the framework for week 2 set up was easier than I expected. In only a few days I had an additional 5 night templates made and ready to be populated with AI and difficulty data. It was precisely at this point where I realized something absolutely earth-shatteringly profound... The game just isn't fun.

How does one make something fun? I don't know, but I can tell you it's a lot harder to do when the core concept of your game is sitting still and looking at things that are out of your control.

I revisted an old idea scrapped from the classic version - a replication of Undertale's FUN system. I thought about it and remembered that a lot of people really like the scripted events in the first game. This leads me to believe that in a FNaF-type game, fun is in part tied to the amount of unexpected things that happen - things that break up the monotony.

• (For anyone living under a rock, Undertale sets a so-called 'fun' variable to a number between 1 and 100 when you start a new game. What number this variable is controls certain events in the game unique to that number, or a range of numbers.)

I made a growing list of 'miniature' scripted events that don't affect gameplay, but are unique compared to other runs. This includes:

• New music

• Rare camera screens

• Minigame changes

• Still coming up with more

A lot of them are pretty subtle, but just noticable enough for you to go "Hey, my game didn't do that" when watching someone else play.

Right now FUN is reset by starting night 1 of the first week. Ahhh does anyone actually read my textwall devlogs? Comment "mouse" if you actually read these, you might get pinned.

The fun events aren't anything major, but they make me smile. Some even contain unused & new music for the game.

Fundamentals, hold the fun

Playtesting the fun values and the in-progress week 2 made me realize something: Even with the fun events, the game still kind of drags. This led to me watching some playthroughs of Mary's Arcade: Classic (And I do actually watch your videos, even if you only have 12 subscribers and 9 views) and observing the play styles.

Mary's Arcade is supposed to be a game where you have to proactively think ahead. The core concept of the game is to reposition characters to ensure they don't arrive at your office before your lure is recharged, yet I gave the players a stalling flashlight that in many cases just prolongs their death for an extra 20 seconds.

People complained a lot about the flashlight, so in early version of Service Pack I made it completely stall all motion so you could not die while it was on, but all that change did was turn

'20 seconds of flashlight hoping that the PA system recharges.'

into

'70 seconds of flashlight hoping that the PA system recharges.'

This completely slowed the game to a halt, or worse, made you fall into a cycle of always barely scraping by with a last-second lure because the flashlight had so much battery. This was killing the game.

My game is supposed to reward proactive thinking and punish the lack of it, so why do I have a giant "PUSH HERE TO NOT DIE" button?

I removed stalling. The flashlight in Service Pack actually serves another purpose on later nights, so I didn't remove the flashlight entirely, but it can no longer stall characters.

This was a big change, but it may be just what I need to get back on the right track. Without being able to stall the game feels a lot more tense; you think about where characters are more often, you look around more.

That's a good start, but without stalling with the flashlight, the only two tools at your disposal are the PA system (lures characters to their next position), and the buzzer (freezes characters on a certain camera) These need to change to work without the flashlight.

Experimental changes

(Everything under this header is WIP and changing almost daily.)

Did you know in early betas of Mary's Arcade, the PA system pulled characters forward and the buzzer pushed them backward instead of freezing them?

I originally intended for there to be a way to push characters back. However, pushing a character back OUT of a room (different from pulling them back TO a room) led to the unintended side effect of being able to push someone back INTO your office after they were already gone.

I hastily removed the push back feature and turned the buzzer into a "freeze" button, and it never occured to me to just make the PA system able to lure in reverse, which is why it can only move characters forward.

Why. Would ANYONE. Ever lure a character closer towards them? They'd just waste a PA system use and then die waiting for it to recharge, but pulling a character back - that could be valuable and used in tandem with the freezing buzzer.

(It also turned out to be really easy to do with the PA system functions I made when porting to GameMaker.)

I'm also experimenting with changes to the recharge times.

I want people to use the PA system for more than just office defense, like de-synchronizing the characters early to make the later night easier and other proactive things like that.

I came up with the idea of the PA recharging faster early on and slowing down as the night goes on to encourage the type of heavier use earlier that I want, but not found a satisfactory way of implementing it yet.

Should recharge times slow depending on the time of the night, or number of uses? Should it slow linearly or multiplicatively? All of these different factors have pros and cons. Either way, with the flashlight being unable to stall, I have found myself taking more action and on more cameras than just the two next to the office.

Conclusion

With all the changes I'm making, the final version of Service Pack should feel rewarding, make you feel smart, and will swiftly punish mistakes instead of just slowing the game to a crawl as you try to stall with a stupid flashlight. Most importantly though, it should feel FUN.

Final note

If you haven't already, go check out Kane Carter's spotlight on Mary's Arcade: Service Pack for a public sneak peak at the models. I also posted some of the new screenshots onto the game page, so be sure to check them out there as well.

• Kane Carter's Game Jolt post

• Kane Carter's Twitter/X post

Or just look below at this horribly compressed version.

Until next time.