Skip to main content

Indie game storeFree gamesFun gamesHorror games
Game developmentAssetsComics
SalesBundles
Jobs
TagsGame Engines
(4 edits) (+1)

It looks very good, the idea is definitely interesting.

I did try playing this one several days earlier, and as soon as it got to the Pick between four twitter responses to that scene in IT I felt strongly that this game was not for me.

However I returned today and decided to power through.

The music starts out nice, and it's subversive like the UI to give happy feelings despite the dark content. However, it's quite repetitive and I ended up muting it around halfway through.

It feels a lot like NationStates in that you have a prompt, then four choices, of which none of them are correct. Except more politically charged than NationStates, somehow. I felt like the choices I made were not at all relevant, and the vast majority of the game is waiting for the analytics on the right to cycle through whatever it was doing, or pretending to be doing.

When interacting with a customer/patient/victim there's no option to do anything other than push the ? button to generate a response, as far as I could tell. In theory this is to see the results of your choices, and in the very first conversation that's true. After that I'm not so sure.

First conversation: Divorced woman, gets a bunch of irrelevant twitter posts and rightfully terminates the call.

Second conversation: Fan complaining that their ship in a fictional story was sunk. Gets generic responses which they don't seem to care about (more on this later)

Third conversation: A cat is lying on a keyboard. Responses irrelevant, it's a cat.

Fourth conversation: It's the second conversation again, but this time... no, it's exactly the same. The generated responses might be different, but they don't affect anything at all, the person said exactly the same things in the same order as before. Is this on purpose, indicating that people just want to hear themselves talk and the other end of the dialogue is irrelevant? Who knows! It proves my choices up to that point had no effect on things, which makes me question why I needed to wait a minute between each decision I made before being allowed to continue.

Then the game is over. I'm left wondering many things.


(+1)

Thanks a lot for the extensive feedback, especially if it didn't hit with you the first time. As these are based on real twitter threads, some of them might be distasteful or unfitting for some. Sadly, I didn't have the time to write anything myself :-/

It's a good feedback that it made you feel like nothing mattered. In general, each choice adds a different data point to your chart and your "AI" is trained on these datapoints:

 

This is the function graph after the first two trainings based on different choices (for speed sake the upper is always picked number 2, the lower always picked number 3). So you can see that the behavior and reaction to the input (x Axis) is completely different between the two. So the development of the patient's values will also be different and based on that also the conversation they have with you.

The waiting time with the patients can be annoying which was the reason I added the fast-forward button. I planned for a more graphical display but given that I had only half the time due to real life, just finishing the core functionality was done ~12 hours before the end... 

In general, the patient's input is based on the three values (it picks only ones that have values close to the current ones) and as your answers change those values, the conversation should differ. Of course, depending on your training they might be similar as well. But what values you get and how they change is based on your decision.


I tried to make this visible with the different text, but I think it shows that I should probably should have put less work into writing text that is 80% similar and rather take a more graphical approach to show that the numbers based on your choices are what matter. In your case, it kinda became like a text autobattler with the feeling that the strategy doesn't matter - which is the opposite of what it should be. 

Thanks for your feedback, it really helps a lot! :)

Thank you for the detailed explanation! It's very interesting to learn the thoughts which went on behind the code. :)

I did miss that there was a fast-forward button, alas. Good to know if I go back and try different choices later.

Is it possible to reach a failure state? Aside from the first day, I always seemed to be doing well according to the boss.

(+1)

In general, each session can have 4 endings:
- All values end up green --> token * high multiplier
- There are no messages that correlate to the current patient values --> no more interest --> token * lower multiplier
- Patient runs out of token --> token * 1
- One of the values goes into the red --> token*0.X (I forgot :D).

In general, you can fail in each session based on your choices during training, but given the current training set, it is easier to not fail. 
The ultimate failure state would be to not have enough money for the bank, which is a bit easer as also a session that ends too fast will allow this.