Artillery prediction fix

[wzdev-ci]

resolution_fixed type_bug | by Adam Olsen <rhamph+warzone@...>

---

The target prediction code assumes all projectiles use a constant speed. Normally this is true, but artillery varies its speed to maintain a ballistic trajectory at longer ranges.

This patch uses sqrt(dist)/30 as a flight time predictor. This is totally a guess and a bodge, but seems fairly effective.

The prediction is still fairly naïve in a number of ways:

• slight turning motions throw it out
• slight hills throw it out, due to speed changes
• despite hills affecting speed, it doesn't actually aim any higher, so it can quite easily miss a unit climbing a steep hill

---

Issue migrated from trac:374 at 2022-04-15 18:08:07 -0700

[wzdev-ci]

Adam Olsen <rhamph+warzone@...> uploaded file warzone-artillery-predict.diff (1.1 KiB)

[wzdev-ci]

Zarel changed status from new to accepted

[wzdev-ci]

Zarel set owner to Zarel

[wzdev-ci]

Zarel commented

---

Yay, math!

[wzdev-ci]

Adam Olsen <rhamph+warzone@...> uploaded file warzone-predict.diff (1.5 KiB)

Handle EMP as well

[wzdev-ci]

Adam Olsen <rhamph+warzone@...> commented

---

New version, handles delays due to EMP. Seems to work for artillery, but I didn't find a convenient example of normal direct weapons having problems to test with.

[wzdev-ci]

Zarel changed status from accepted to closed

[wzdev-ci]

Zarel set resolution to fixed

[wzdev-ci]

Zarel commented

---

Committed in [7155].
Backported to 2.2 in [7156].

[wzdev-ci]

Adam Olsen <rhamph+warzone@...> commented

---

Indirect fire has four range categories (for trajectory purposes, not hit chance purposes):

Long: fixed angle, variable (high) velocity
Medium: low variable angle, fixed velocity
Short: super high variable angle, fixed velocity
Point blank: direct fire

My previous patch confused the point blank/short threshold for the medium/long threshold. My new patch fixes this, although the threshold itself is a little fuzzy.

The original code had correct prediction for point blank and medium. My previous patch added correct prediction for long. However, there is still no correct prediction for short. Short also has significant problems with height differences, so it is often unable to hit targets that are much lower than it. I've added an extra range check (psStats->minRange < TILE_UNITS*3) to reduce how often it gets used, but I've been unable to eliminate it (without making point blank excessively large.)

IMO, there's a lot of ugliness here, but I want to get the medium/long threshold fixed before getting too deep in fixing other things. I did add a small check for projectiles passing below the map, but that's usually because they skipped past their real target.

[wzdev-ci]

Adam Olsen <rhamph+warzone@...> uploaded file warzone-predict-3.diff (3.0 KiB)

[wzdev-ci]

Zarel changed status from closed to reopened

[wzdev-ci]

Zarel changed resolution from fixed to ``

[wzdev-ci]

Rhamphoryncus uploaded file warzone-predict-3b.diff (57.5 KiB)

Missed a bit. Bluh.

[wzdev-ci]

Per commented

---

That last patch is full of noise.