So let me just start off by saying I really, thoroughly enjoyed my playthroughs of GDT. However, like a lot of people here I found myself looking for mods to extend my enjoyment with the game. I then ran into the Wiki as well as the success guide/review algorithm to learn more about the game and how to make successful games - http://gamedevtycoon.wikia.com/wiki/Success_Guide
This text in particular really bothered me in regards to the review algorithm:
"This means that you compete only against your own “Top Score” (or a value of 20, until you make a Top Score). Your first Top Score is established when you release a game which has a Review Score of 9+ (a bit more complicated than that actually, but let’s ignore that for a moment). When you release a 9+ game, your Game Score of that game is assigned as a new Top Score.
Therefore, in order to get a good Review Score, you need to improve over your best previous Game Score.An optimal strategy would therefore be to increase your Game Score by ~11% with each release, and thus get near-perfect review scores every time. However, increasing your Game Score by more than 20% is harmful for the future, because it is effectively wasting an opportunity for getting good reviews."
So I guess my question is - is anyone working on a mod that has a better system for calculating the review score? I find that I literally cannot play my old saves anymore (in Y30 or 40 or whatever) knowing that my game scores are now toast after scoring a perfect 10 game.
Thoughts? Maybe I’m way off base here? Someone lead me back to the light
Or, is anyone working on a mod with a new review algorithm?
First of all, a large part of the wiki is from version 1.3.9 and does not take the updated review mechanics into account. In version 1.4 we largely addressed these issues. If you create a great game then you won’t get bad scores just because you created a perfect 10 before.
The idea that creating a great scoring game will make future games more difficult is still true however. The effect isn’t immediately though. After you release your super hit, the market will take a bit to adjust its expectations but this is always within limits and I think reflects a logical progression in difficulty.
Having said all that, it’s entirely possible to create a different review algorithm through mods, if anyone is interested in that.
I am trying to find ways to improve the algorithm. This will be a work in progress for quite a while tho, since i am currently engaed in 3 other projects AND have university to attent to as well.
Goal is to improve the algorithm to feel more natural and punish less for making good games. Holding back to make the most money just does not feel right, but i understand the thoughts behind the current implementation as well.
Current Ideas include:
having a game difficulty slider. Each having a fixed targeting score
for games.
using a mix of fixed targetscore and your own best games
to easen the curve a bit without using spreatsheets
having a fixed and pretty high target score and in return also have % based
advancement costs to make it easier if falling behind the curve and
harder if you get to far ahead.
Feel free to add your own. Right now id prefer the 3rd option, maybe combined with the second. After all it IS realistic that ones own games adjust the expectations of future games. Also some great ideas came from PatrickKlug himself: Makeing a scoreboard ingame with the top 10 games, simulate competition and making it feel more like a race against said competition. I think this should be included as well, ontop of the 3 points above.
But all this will take time and i am unsure when i can start. Been looking over the code for some and pretty sure it is doable tho.
Hmmm, I was thinking - is there anywhere/resource where I can find an explanation of the new/updated Algorithm?
I’m sure someone far smarter than me can look through the code, and I’m not asking anyone here to do that necessarily but if someone has already done that and made it available somewhere I’d be interested in taking a look.
Sadly, I never got around to that.
I looked at the Code back then (They where generous enough to let me see a copy of the non obfuscated code even!), but I only had very limited time. It looked like the algorithm was not so much part of a single routine, but more like backed into several steps and sub steps along the way. It was not easily noticeable what and where to change to make the wanted changes.
Sadly that is about all I remember from 6 years ago. There are topics about this every now and then on the steam forums as well, but I guess I was not alone in having issues actually pulling it off, since no mod ever came out that fixed the review system to my knowledge.