Poll: Which AI modifiers/plots do you use?

[Atomikkrab]

Yea, it isn't so much the baseline difficulty (schizo is usually higher there) but the peak difficulty.

It isn't the 100 waves you beat that kills you (except for econ-death-spiral situations), it's the 1 that punches through because it was all-bomber, max-time, double-wave, at-the-same-time-as-a-CPA-and-an-exo, etc.

No I beat that, even with the EMP guardian shutting down everything and costing me my black hole machine and most of my spire fleet from the EMP taking down their modules.

I probably wouldn't even lose the bhm with those super fortress upgrades...

[Draco18s]

No, I'm not saying roles would make them easier than Schizo.

Uh...

Schizo waves would be more interesting if they only picked 2 or 3 ship types to make the wave.  That would make any two non-mono waves different, and create some much nastier combos when the AI has stuff like Munitions, Shield Bearer and such.

This is actually a really nice idea. But rather than 2 or 3 ship types, how about 2 or 3 ship "roles"?

The whole roles vs. types is about schizo waves so I don't see how that change would make schizo waves easier than schizo waves.

[Hearteater]

You appear to be very confused.  I'm afraid I'm at a loss on how to correct this.

[Draco18s]

You appear to be very confused.  I'm afraid I'm at a loss on how to correct this.

1) Changing how schizo waves are determined HAS NO EFFECT on the difficulty of non-schizo waves for ANY AI type (in fact, it is mathematically impossible for a change in schizo waves to alter the objective difficulty of an unrelated game mechanic).
2) Making schizo waves harder is the whole point.

[Hearteater]

You have just confirmed you are massively confused.  Why you would respond with either of those two point, to me specifically, in this thread specifically, is beyond all reason.  Start here in this thread were I suggest using ship types, and see later down the page TechSY730 suggesting using roles in stead of types.  Then continue as I explain how using roles is less effective at making Schizo waves harder than using types.

[Draco18s]

You have just confirmed you are massively confused.  Why you would respond with either of those two point, to me specifically, in this thread specifically, is beyond all reason.  Start here in this thread were I suggest using ship types, and see later down the page TechSY730 suggesting using roles in stead of types.  Then continue as I explain how using roles is less effective at making Schizo waves harder than using types.

And then see my rebuttal.

I don't know that I'd go to roles.  Part of the danger that limited ship types can present in a wave is you not having a counter for one of them.  It probably wouldn't matter for a lot of AI Types, but some that have a lot of free ship types unlocked would feel pretty similar to the current Schizo, ironically making only harder AI Types easier.

This doesn't make sense.

If the AI is limited to two ship TYPES, it throws....

Z Chameleons
Frigates

Odds of me not having something that counters the whole wave: very very low (I need exactly two units: fighter and frigate)

If the AI is limited to two ship ROLES, it throws....

Bombers (sends bombers and z chameleons)
Artillery (sends frigates and grenade launcher)

Odds of my not having something that counters the whole wave: much higher.  Sure, the two artillery units both have the same armor type, but I wouldn't be able to send my fighters in to attack the bombers without them getting beat up by the frigates and grenadiers.  But I wouldn't be able to send in my AOE immune frigates to engage the grenadiers because of the bombers.

Then you asked me if schizo are easier or harder...

If you think schizo are easier, then you should realize using [roles over types] pushes [schizo waves] towards being easier.

Without explaining how that's true.

Unless that post is actually saying,

If you think schizo are easier, then you should realize using [roles over types for schizo waves] pushes [non-schizo waves] towards being easier.

Which makes no sense.
(Brackets added to group modifying phrases)

[TechSY730]

Limiting to only a few ship types will increase diversity between one wave and the next (on average).
Limiting to only a few ship roles will increase diversity within one wave (on average), even if it is a little bit more likely that one wave will sort of look like another.


Too far in one direction (mono-waves) leads to very inconsistent difficulty depending on how much chosen works as a standalone offensive/"take out the planet", with a huge gulf between the min and max difficulties.
Too far in the other direction (the fully fixed nature of the current scizo) leads to far more consistency, but consistent at a lower point than the max of the mono-waves.

I was trying to compromise some. While selecting only a few ship types seems like it will suffer, but to a lesser degree, this inconsistency, and make waves a tad bit too un-diverse within a single wave. That's why I chose roles. (Choosing roles would also help out in cases where the AI would choose 3 support ship types or 3 anti-light mobile stuff ships, ships while are fine, don't work so well on "take out the planet" offense in a standalone capacity. Choosing roles would make these sorts of "worst case" choice less likely, without "cutting down" the "practical" max difficulty by too much)

[Hearteater]

And then see my rebuttal.

Which was terrible.  It consisted of one example, both of whose ships are countered by triangle ships.  I ignored it because there was nothing to argue about.  Of course you can find a possible example that's easier than a Schizo wave is presently.  Just like a Schizo wave can be tougher than a mono wave.  No one plans there defense around the easy waves, they need to stop the hard ones.

Then you asked me if schizo are easier or harder...

So I skipped wasting time arguing about nothing, and wanted to determine if you even understood that the current Schizo waves were easier.  Which clearly caused confusion.  It has already been explained why Schizo waves are easier in this thread, more than once, by myself, TechSY730 and Keith.  At minimum.  I may be forgetting someone.  So if you don't understand that, just back up and check that out.

Once you understand WHY Schizo waves are easier, you should understand why I said this (broken into two separate points):

1) then you should realize using roles over types pushes [a new Schizo wave design] more towards the current, easier, schizo waves,
2) and worse, it does so mainly for the harder AI Types.

Point 1 simply comes from using roles instead of types diluting the wave, making it more like the current Schizo waves.

Point 2 acknowledges that since not all AI types have all ship types, the contents of a role will vary by type, possibly being only a single ship type.  Which means for some AI types with few unlocks, roles will be the same as types.  Specifically, easier AI types.  This changes the behaviour of the proposed Schizo waves in a way inverse to AI type difficulty, which is counter-intuitive.

I was trying to compromise some.

When you compromise between a good solution and a bad solution, you still have a bad solution.

[Draco18s]

Of course you can find a possible example that's easier than a Schizo wave is presently.

Ahem.  My example was where it was more difficult not less.

It consisted of one example, both of whose ships are countered by triangle ships.

...

Don't the triangle ships cover all hull types?

So I skipped wasting time arguing about nothing, and wanted to determine if you even understood that the current Schizo waves were easier.  Which clearly caused confusion.  It has already been explained why Schizo waves are easier in this thread, more than once, by myself, TechSY730 and Keith.  At minimum.  I may be forgetting someone.  So if you don't understand that, just back up and check that out.

...

Schizo waves aren't harder, I never said they were.

You don't even read my posts, do you?

[TechSY730]

I was trying to compromise some.

When you compromise between a good solution and a bad solution, you still have a bad solution.

Just to clarify my goals.

My goals are to keep up schizo wave composition variety within a single wave, bump up average difficulty a good deal, maintain some consistency of difficulty between different waves, and bring in just enough of what makes mono-waves have such a high max difficulty without compromising variety within a single wave much.

Both of our ideas would bump up "practical" min difficulty, bump up "practical" max difficulty, bump up average difficulty a little, and keep pretty decent within a single wave variety.

My idea (using roles), compared to yours (using types), tries to bump up the min difficulty a bit more and keep up some within a single wave variety some, at the expense of not bumping up "practical" max difficulty as much. How average difficulty would compare between the two ideas is rather tricky to answer (see below)

It really depends on which attribute we want to maximize which one is the better choice.

Just to clarify, BOTH ideas would raise the "practical" max and average difficulties, so both ideas would make schizo waves more threatening. It's how much we care about the other variables that will influence the decision (and how much we care about smallish differences to the max difficulty, and the yet to be determined effects on max-difficulty)

However, there is a fundamental question:

Is, on average, (assume ratios have already adjusted for ship caps) how does the difficulty, the expected "practical" min, average, and "practical" max difficulties vary when facing 1/N ratio of N types of ships? (Note, the exact value will vary strongly with which types are used. Complicated matters is that the ship type selection are NOT independent events, especially in the "role" selection idea, making describing the expected distribution of these types even trickier.)
The issue is that the functions are not non-decreasing and not non-increasing. For example, average difficulty may go up from N=1 to 2, but go down from N=1 to 10, but stay the same from N=9 to 10 (not sure if any of these are true, but this is just an example of the kinds of effects that can crop up). This makes the shape of the curves complex, and we need to get a rough idea of where the "peaks and valleys" are in this curve before we can compare what the difference between our two ideas would be in the average case.
Plus, this assumes equal distribution (scaled for ship cap), which does not have to be true (like a bomber focused schizo wave, for instance), making things even more complicated.

[TechSY730]

Hey, just thought of another idea.

Well, if we do use the "choose N ship types" idea, I realized there is another place in the game that does this, planetary reinforcements.

There is a good chance that a planet will pick three "focus" types of ships for its defenses. Sometimes, it might choose them independently. Other times, it might try to choose ship types based on their roles. If it does pick based on roles, it will still pick 3 exact ship types, but the selection of those types may consider their of roles. (keep in mind, even if it does select 3, there is still a small chance that a any ship type will picked, even if not on that list of 3. I think it was 5%-10% per defender of this happening, so you will still see some small smattering of other defending ships around on planets that chose "focus types")

Why not that for schizo waves. It would pick N ship types, sometimes independently/randomly, but other times using that same "role based consideration logic" to find which types to fill those N slots it gets. (maybe with or without that <=10% (chosen on a per-wave basis, but not to exceed 10%) or so portion of the wave that is free to chose whatever)

[Hearteater]

You don't even read my posts, do you?

Every one of them.  But it has become clear you just aren't understanding.  Whether that is an issue on your end in comprehension, or mine in communication, I'm uncertain.  But I have decided it isn't worth it to continue to try and explain this to you.

[keith.lamothe]

It's Hearteater vs Draco18s, on pay-per-view! (oh, wait, the forums are free, hmm).


Anyway, the principle is very simple: a wave split across, say, 5 distinct ship types will almost always be less dangerous than a wave split across 1 or 2.

There are a few exceptions, but encountering them randomly would be a rare event.

Now, if "by role" meant:

"pick two random roles, then pick one type from role 1, and pick one type from role 2"

instead of

"pick two random roles and include all types from both roles",

that would probably be at least as dangerous as:

"pick two random types"

And perhaps moreso, but probably not in the general case.

[TechSY730]

"pick two random roles, then pick one type from role 1, and pick one type from role 2"

My second form of the suggestion (which I just posted), which if I understand correctly, the planetary defender "focus" types have a chance of doing this sort of thing, right? (Though it has a chance of picking the same role for ship 1 and ship 2, for example). Also, it could be up to 4 roles.

"pick two random roles and include all types from both roles",

My first form of the suggestion. (again, up to 4 roles)

[Draco18s]

You don't even read my posts, do you?

Every one of them.  But it has become clear you just aren't understanding.  Whether that is an issue on your end in comprehension, or mine in communication, I'm uncertain.  But I have decided it isn't worth it to continue to try and explain this to you.

It's likely your end, as I've stated multiple times that schizo waves are easier than mono waves, and you keep implying the reverse.