Here is a fact about the games industry that ought to be strange.
Souls games are, by reputation and largely in truth, the least accommodating
mainstream games there are. They kill you constantly, explain almost nothing,
and take the better part of a hundred hours. One boss can take fifteen or
twenty attempts. Nothing in that description sounds like a game people finish.
So the puzzle is: why does the harder game keep the player that a gentler one
lost?
The answer is the thing you would least expect, and it is the whole lecture:
because the boss is hard, you die in about two minutes. You therefore find
out you were wrong two minutes after starting, and you can be back in front of
it fifteen seconds later. The difficulty is not an obstacle the game overcomes
on its way to being enjoyable. The difficulty is what produces the short
loop.
Time to failure — how long from starting an attempt to finding out it
didn’t work.
Time to retry — how long from that until you can try again.
A Souls boss: ninety seconds and fifteen seconds. You die a minute and a half
in, you are back at the fog gate a quarter of a minute later.
A fighting game combo: you go into practice mode and drill a sequence whose real
payoff arrives in a match several sessions from now. Time to failure is short;
time to finding out whether it mattered is days.
An assignment: three weeks to failure, and no retry at all.
The Souls loop is short because the boss is hard. Those are not two facts
about the game, they are one fact stated twice: an enemy that can kill you in
two minutes is an enemy that can tell you that you were wrong in two
minutes.
A ninety-second attempt is short enough that you only retain one or two things
from it. It always sweeps after the second swing. That is a small, cheap,
completely testable hypothesis, and the interesting part is what it does to you:
you want to know right now whether that one thing is enough.
It usually isn’t. So you learn a second thing. Twenty attempts later you have
twenty small true statements about this boss and you win, and at no point did
you have to decide to be disciplined.
Curiosity carried you, and curiosity is much cheaper than discipline. This is
the design principle the whole second half of the course is built on: make the
lesson per attempt small enough that testing it is the obvious next move.
Legible tells. The boss telegraphs. You lost because of something you can
name, which is what makes the lesson extractable. A game that killed you for
reasons you cannot reconstruct produces no hypotheses, and it is the failure to
be legible — not the failure to be fair — that makes people quit.
The boss does not change. Between attempt four and attempt twenty the only
variable that moved is you. That is a rare and precious property, and most of
your studying does not have it.
Have something installed and ready. Any of the Souls-likes work; you want a game
with discrete boss encounters and a short walk back, not an open-world game you
can wander away from when a fight goes badly. Wandering away is exactly the
behaviour we are studying, and this week you need to not have the option.
The slides, as text
Every slide in order, with a description of each figure. Generated from the deck source, so it cannot fall behind it.
1
The hardest game is the one people finish
SLOP3733 · Week 6
FigureA fog gate: a stone archway with gold mist standing in it
2
By every rule, this should not have worked
A Souls game kills you constantly. It explains almost nothing. It takes about a hundred hours.
One boss can take you fifteen or twenty attempts before you get past it.
Nothing about that sounds like a game people finish.
3
Because the boss is hard, you die quickly
A minute or two in, you are dead.
So you find out you were wrong a minute or two after starting, and you can be back in front of it fifteen seconds later.
Nothing else in the game is doing that. The difficulty is.
4
Being hard is what makes the feedback fast
It is the cause, not an obstacle the game has to get past
5
So measure the loop
FigureA broken ring of tick marks closing on itself, with a short bright arc where the loop restarts
6
Two numbers you can time with a stopwatch
Time to failure — you start an attempt, and you find out it did not
work.
Time to retry — from that moment until you can be trying again.
Everything has both. Reading a chapter has both. An assignment has both.
7
The three loops, drawn to scale
FigureThree feedback loops on a logarithmic time axis from one second to one month. A Souls boss: ninety seconds to failure, fifteen seconds back. A combo in the lab: three seconds to failure, but about two days before a real match tells you whether it mattered. An assignment: three weeks to failure, and the retry line runs off the axis because there is none.
8
The harshest one has the shortest loop
And it is harsh and short for the same reason
9
Now what happens inside those two minutes
FigureForty faint horizontal lines and one bright gold line among them: everything a long session produced, against the one thing short enough to keep
10
Two minutes is only long enough to remember one thing
Not five. You come out of that fight holding one observation, maybe two, because that is all a two-minute attempt leaves you room to keep.
11
It always sweeps after the second swing.
That is the one thing. It cost you two minutes, and you can check it immediately.
12
"If I fix just that one thing, do I get through?"
13
You are already clicking
You did not decide to keep going. You did not talk yourself into another attempt. You wanted to know the answer to that question, and the only way to find out is to walk back in.
Curiosity got you to attempt twenty. Nothing had to be overcome.
14
Twenty attempts, one thing learned each time
FigureTwenty attempts at one boss as a bar chart. Each bar is how far into the fight you got, rising unevenly from under a tenth to the kill on attempt twenty. A gold dot sits above every bar: the single lesson that attempt was short enough to carry. The whole row of twenty costs about thirty-five minutes.
15
Compare that with a fighting game
A fighting game asks you to love it first. You go to training mode, you grind a combo against a dummy, you learn frame data — and you find out days later, in a real match, whether any of it was worth the hours.
That works, but only for people who already love the game enough to do unrewarded work.
A Souls game asks for something everybody has. Wanting to know what happens if you fix one thing.
FigureA training-mode dummy above a five-input combo string, with a three-frame link window marked on a one-second ruler of sixty ticks
16
Two things had to be true for that to work
You could name what killed you. The game showed you the sweep. If you cannot name the cause, there is no one thing to take into the next attempt.
The boss never changed. Between attempt 4 and attempt 20 the only thing that moved was you, so getting better was something you could feel.
17
Neither one is true when you study
You cannot say what went wrong in the last three hours. Nothing showed you the sweep.
And the material is not the same thing twice. Every time you come back to it you are reading something slightly different from what you remember.
So you get the punishment without any of the things that made it bearable.
FigureA page under a lamp whose lines are illegible and printed twice, slightly out of register, beside a cold mug and a clock with no hands
18
This week you play
One boss. Twenty deaths. One written line each.
FigureTwenty tally marks scratched into dark stone in four groups of five, counting deaths rather than wins
19
What to bring back
Twenty lines, one per death. Each line has to be a claim about the fight that could turn out to be wrong.
It always sweeps after the second swing is a claim. Attempt twenty-one either confirms it or it does not. I was bad at it is not a claim, because no attempt settles it either way.
Bring the log to the Lab in week 7. There is nothing to do in that Lab with an empty one.