Draygon numbers

[osse101]

• Charge without plasma is limited in how much damage you are really doing
• Supers are {opener:5, early_swoop:2, goop:5, late_swoop:0} so instantly ending the fight with 20 supers is incorrect
• I don't believe you can farm goops and dodge draygon in Hard. There are random drops, but not enough to refill a large number of missiles, which is what I wanted to change.
• Added more interactions with Suitless & Morphless. I find I get hit constantly when morphless, and suitless+morph takes a ton of practice. And they take a similar rate of random hits.
• Without Gravity I found I could fit half as many missiles during a goop, and basically skipped the first wave of goop farms per cycle.

Have a look, see what you think.

[blkerby]

The changes move in a good direction. I did another pass now to try to refine some things a little more:

• Cut firing rate in half without Gravity: I believe 0.5 * 0.3 + 0.7 * proficiency was intended to be 0.5 * (0.3 + 0.7 * proficiency).
• Fix time computation to be supers_used / super_firing_rate instead of supers_used * super_firing_rate.
• For farm proficiency, rather than having a discontinuity at proficiency = 0.6, add a quadratic term to implement a similar idea, making the farm proficiency stay relatively flat at low difficulties and rise more sharply at the higher difficulties.
• Avoid decreasing the base firing rates quite so much, but make them also vary based on having Morph or not.
• Avoid making Charge hard-require Plasma or a proficiency threshold. But change the computation so that Charge is only an alternative to farming Missiles, instead of stacking damage on top of Missiles.
• For the Gravity + Morph scenario, avoid reducing the Basic damage to 0.25 hits per cycle, as that would need higher skill than we expect for new players. So keep it at 0.5. The aim is to make the fight require 6-7 tanks with Morph, double suit, and plenty of ammo, to roughly align with what the vanilla game expects.
• For the suitless morphless scenario, avoid having a rate of damage of more than 1 hit per cycle. Instead apply a transformation to make the damage not scale down as much until the higher difficulties.
• For the other scenarios (Gravity and/or Morph available), do the opposite, by making the damage rate fall off more quickly at lower difficulties. Since we are nerfing the assumed firing rate and farm rate quite significantly, this was necessary to keep the expected damage in a reasonable place for Expert difficulties and higher. And I think it reflects the reality that basic dodging techniques can be learned fairly quickly. It could possibly use more tuning still.
• Model patience tech more completely/accurately. Again, because of the changes to the firing rate, it is now more likely that a strat goes out of logic due to patience. It may still need more tuning to get it into an ideal place.

[blkerby]

Worth keeping in mind that the suitless morphless variant is not going to come into logic without a change on the sm-json-data side. I assume we may want a notable for it. A pause abuse version of the notable could be interesting too, though I imagine that could get complicated; to avoid a blowup in possible combinations of weapons and/or patience, we probably would still need to use the randomizer's model, e.g. using a draygonKill requirement with a parameter to indicate the pause abuse, similar to what we did with Ridley to handle the G-Mode CF stuck situation.

[blkerby]

It will be interesting to see where the limits of the suitless morphless version turn out to be. I think this initial model is fairly conservative in assuming a lot of damage even at the highest difficulty. I had some success with doing low spin jumps to avoid swoop damage; hanging out more on the right side of the room seemed to help. I had a hard time dealing with the goop phase but I imagine that's just lack of familiarity with the suitless fight, since I don't imagine Morph would necessarily help with the goop phase.

[osse101]

Suitless Morphless:

• With pause abuse its not bad, around Extreme+.
• I could also pull off my ammo 1 cycle kill without a pause abuse, but the goops are still a little random.
• I don't know if either can or should be modelled, but could be notable.

Review:

• My first pass didn't scale great into Very hard, so keeping a fluid algorithm is good.
• I saw from your second pass that missiles could be farmed too much, weak charge was being used, and strong charge was performing the fight worse when adding more supers and missiles.
• 180s fight can feel pretty bad still if its all dodging and farming. When I did it with 5 missiles and otherwise good items, it took 4 minutes including entry/exit. Its hard to know what feels right because when we do the low equipment fight it feels like it goes on for ages.

This commit:

• Capped missiles per cycle for low capcity situations
• Dumb check that you can take damage in no suit + no tank situations. When the energy cost is just under a hit threshold its going to feel scary, still.
• Weakened value of missiles & missile farm
• Lowered DPS of plasmaless charge a bit

Gravityless looks like it could be too relaxed now.

[blkerby]

• Adjusted the energy capacity check (to be able to survive a hit) to not make it a hard requirement for doing the fight. Instead, made it so that it just assumes no energy farming is possible if you can't survive a hit. This reflects that the fight can still be possible if you have good enough weapons and dodging to finish it before you would be expected to take a hit.
• I'm not sure weakening Missiles this much at the lower difficulties makes sense. On Basic this would put Missiles accuracy to 16% compared to 40% for Supers. I would think if anything it should be the other way around: at low difficulty, you can spam Missiles to clear the goops and get some hits onto Draygon, whereas with Supers a higher fraction of the shots will get eaten by the goops. Maybe it would make more sense to just lower the base firing rate for Missiles, or increase it for Supers? Keeping in mind the base rates are for the ideal scenario where you have Gravity + Morph, with the other scenarios being scaled down from there. Either way, if we vary the accuracy by weapon, then we need to adjust how good_time is calculated for the patience requirements. I had been thinking of maybe breaking out a separate function to compute (energy, time) as a function of proficiency; then we can just call it twice, once with the actual proficiency to get the energy required, and once with a proficiency of 1, to get the good-execution time for patience requirements.

[osse101]

I lowered patience instead of accuracy. I like the numbers throughout now. Basic has funny numbers 😸