How I balanced Mythos Slayer: 2.42x per boss, and a curve I stole from my own game

I built this game. Every number below comes from my own build and save data, not from playing someone else's clicker.

The hardest part of making an idle clicker is not the code. It is one number.

In Mythos Slayer, boss HP multiplies by 2.42 every time you advance. Boss 1 has 300 HP. Boss 6 has 24,900. That is 300 multiplied by 2.42 five times over.

Getting to 2.42 took longer than building everything else in the game. This is the record of how I got there, including the version that collapsed, and the problem I looked at and decided not to fix.

Version one scaled by digits, and it broke immediately

My first approach was the obvious one. Give each boss an HP value one digit longer than the last. Hundreds, then thousands, then tens of thousands. Clean to implement, easy to reason about.

It fell apart in the early game.

Scaling by digits means every step multiplies by ten. In an idle game your damage does not grow by ten per step. It grows in stages, gated by which units you have summoned and how far you have upgraded them. So the gap between what the boss demanded and what a player could actually produce opened up almost immediately.

The inflation was not gradual. It hit within the first few bosses, which is exactly where you can least afford to lose people.

The problem was that I could no longer feel my own opening

Before I could fix the curve I had to admit something uncomfortable. I was the worst possible person to judge my own early game.

I had played that opening so many times that it had stopped registering. I knew where every upgrade sat. I bought things in the right order without deciding to. A first-time player has none of that, and I had no way to see the game through their hands anymore.

So I stopped playing it myself and started running automated playthroughs instead, with no knowledge of the optimal order baked in. They bought the wrong thing. They sat on gold they should have spent. They stalled in places I had long stopped noticing, because I had been routing around those places on instinct for weeks.

Several separate runs jammed at the same point. That was what I needed. It told me the wall was real and not just me second-guessing my own design.

It did not hand me the answer. It gave me the shape of the problem, which is the part you cannot fix until you can see it.

The fix: I did not design a new curve. I ported one

This is the part I would probably dress up if I were trying to sound clever, so let me just say it plainly.

I did not derive a fresh difficulty curve for Mythos Slayer. I took the one I had already tuned for Soldier Typing, my typing game, and brought it across.

That curve was built by working backwards from average human typing speed. Take how fast a typical person actually inputs, work out roughly how much damage that produces at a given stage of progression, then set boss HP so the fight lands inside the window you want. Not a multiplier picked because it looked right. A multiplier that falls out of a target time-to-kill.

The translation to a clicker was one substitution. A fast typist becomes a fast tapper. Both are just input rate, and the curve does not care which key you are hitting.

The honest reason I did it this way is that reusing something that already works is easier than building it again. But it turned out to be the better call regardless, because that curve had already survived contact with real players in another game. Starting from something tested beat starting from a blank sheet.

What came out the other side was 2.42.

The full boss HP curve

Boss 1 and Boss 6 are read directly off my own save. Everything between and after is the 2.42 multiplier applied forward. If you want to check my working: 300 multiplied by 2.42 five times is 24,899.7, and the game displays 24.9K.

Does it still hold twenty-seven bosses later?

It does, and I can show you.

Boss 28 has 6.92 trillion HP. Run the same multiplier out twenty-seven steps from Boss 1 and you get 300 multiplied by 2.42 to the twenty-seventh, which is 6.92 trillion.

That is the whole point of picking a single ratio instead of hand-authoring each boss. One number, applied consistently, and the curve is still doing exactly what I designed it to do twenty-seven bosses later. If I had tuned each fight by hand, that consistency would have drifted apart somewhere around boss ten and I would never have noticed.

The other two numbers in that screenshot matter as well. Twenty-one units on the field, up from ten at Boss 6. And two rebirths completed, which is what made reaching Boss 28 possible at all. More on that below.

Two systems, not one

Damage in Mythos Slayer comes from two places, and understanding the split is most of the game.

Summon puts units on the field. They deal damage on their own, forever, whether or not you are touching the keyboard. Farmer and Villager do not fight at all, they collect gold.

Upgrade is everything else. Tap damage, crit, gold rate, attack speed. These are global and they never stop mattering.

Here is the Upgrade panel at the start of a run.

The cost order is the design telling you something. Kitchen Knife at 60 G is what carries you through Boss 1 before you own a single unit. Gold Rush at 1.20K is twenty times that, and it only pays off once you are tapping a lot. The gap between them is roughly where your first summons should be landing.

What the unit tiers actually look like

Real save data from Boss 6. Ten units summoned, 1.69K DPS, 316K gold per minute.

Two gold generators sit alongside these. At Boss 6 my Farmer is level 12 collecting 741 G, and Villager is level 0 collecting 57 G. At the very start of a run, Farmer costs 25 G and collects 1 G.

Read the cost column against the DPS column and the stepping is obvious. Swordsman to Heavy Infantry roughly doubles your damage for twenty-three times the price. Wizard to Knight roughly doubles it again for 3.6 times the price. Knight to Paladin buys 1.5 times the damage for 3.6 times the price.

That is the shape I wanted. Each tier is a real jump, and each jump costs enough that you have to commit to it.

How to actually play this

• Buy gold generation before damage in the early bosses. Damage you cannot afford to upgrade is worth less than income you can. Summoning a single Farmer takes gold per minute from 36 to 126 for 25 G, and that compounds into everything else.
• A level 0 unit one tier up usually beats a heavily levelled unit one tier down. Heavy Infantry at level 0 already out-damages Swordsman at level 2. If you can afford the next tier, buy it instead of levelling what you have.
• The Wizard to Knight cost jump is where most runs stall. 94.6K to 342K is the largest relative wall in the mid game. That is the point to consider resetting rather than grinding through it.

Why Rebirth exists: the screen is a hard limit

Rebirth, the prestige system, was not in the original plan. It exists because of a constraint I could not design around.

Your damage comes from units standing on the field, which means the number you can have is bounded by how many fit on the screen. That is a physical limit, not a balance decision, and no amount of tuning removes it.

On top of that, units step rather than scale. Farmer, Swordsman, Dwarf Warrior, Archer, up through Knight and Paladin. You do not smoothly ramp. You unlock a tier and then sit at it until you can afford the next.

Put those together and the late game has a ceiling. Bounded unit count, stepped damage tiers, boss HP climbing at 2.42 per step. Playing my own build far enough, I hit a point where the boss curve kept going and my damage had nowhere left to go. The run was stuck.

Rebirth is the release valve. It unlocks at Boss 10, at 854K HP, which is roughly where that ceiling starts to bite. Reset, gain a permanent multiplier, come back with a higher floor.

I rebuilt the Rebirth multipliers once as well. In the first version, resetting the instant it unlocked was always correct, which made the choice meaningless. Now the reward scales with how far you got before resetting, so holding out actually pays.

The early-game churn I decided not to fix

When I looked at where players were leaving, the answer was blunt. The biggest drop-off was in the early game.

The obvious response is to soften the opening. Cheaper first upgrades, a faster Boss 1, a smoother on-ramp. I did not do it.

The payoff of an idle clicker comes from contrast. The opening is supposed to feel tight. What makes the late game land is the memory of that grind, and then the moment your numbers run away from you and a boss that should have taken minutes evaporates. Flatten the opening and you flatten the payoff with it.

So I made the trade knowingly. I accepted losing players at the start in exchange for the people who stay getting the thing this genre is actually good at.

I am not claiming that is right for every game. It is a genre call, and I made it with the numbers in front of me rather than by ignoring them.

What I am not adding

The thing I keep wanting to build is a skill tree. It would fit, and it would give the Rebirth loop more to chew on.

I am not going to add it.

The appeal here is that you open the page and understand the whole game in about ten seconds. Summon, upgrade, push, rebirth. A skill tree adds a layer you have to study before you can play, and for a browser game someone opens on a break, that cost is higher than the depth is worth. Stacking on systems is the easiest way to make a simple game worse.

What this cost to build

Honestly, Mythos Slayer was one of my cheaper projects. Almost all the effort went into the balance work above rather than into implementation, and the balance work itself got shortened by lifting a curve I had already tuned elsewhere. The games I built afterwards, particularly the fishing game and the Unity projects, took far more time and far more of everything else.

Which is its own lesson. The expensive part of this game was never building it. It was deciding what one number should be, and then being willing to throw away the first answer.

Mythos Slayer runs in the browser in English, Japanese and Chinese. Free, no install, no account.