That's not entirely how this works. It's all based on analytics, statistical probabilty of a real user spending money inside the game and levels of money that user is willing to spend.
Players ability to spend money:
A. Are you playing on a LG G5 ancient Android phone, is your IP address in Wyoming USA or Canada (currency conversion)
or
B. Are you playing from an iPhone 13 Pro Max from a New York USA IP address
Player Engagement:
A. Did you play 700 battles to get an achievments and play daily
or
B. Do you barely log in abandoning the game for a week to do other things.
Are you good at gameplay:
A. Do you have a strategy you've worked out to win well with your current mech load
or
B. Are you constantly getting blasted
Are you frugal:
A. Do you have a high balance of game dollars or a-coins at the moment
or
B. Are you nearly at 0 on all of them
All these metrics are taken into account - collated, compared with other players data to predict behaviour, this is what controls what offers, archivement sprees, development durbies and other things you'll be exposed to, the difficuilty setting to your AI opponents and etc.
Also all these are piloted on a select group of real players, just like facebook and google product design rollouts. So they'll demo this in the India market, and then launch it worldwide if Player Engagement metrics go up. The biggest goal for this game design is usually to keep you in the game as long as possible and to open the app as often as possible. It's called attention based economy, he who captures your attention and time will likely get your money.
So when anyone from Plarium speaks about re-balancing, what they relally mean is a fine balancing line between Player Engagement (numbers of concurrent players and total hours played globaly) and Sales of paid items inside the game.
E.G. If all your achivements and goals were easy to attain because of fair gameplay, then you'd play - get your XP/Coins/Accolade Bonus and log right out till the next game reset cycle, that would be counterproductive for the business and if the trend is that many people do this, then obviously the Ai difficuilty is tuned up to the next level or parameters are tweaked till the user engagement numbers go up again.