Scalar/Version History

Scalar Sub-pages:

Scalar - Version History - Understanding Scalar

---

• 0.01e22 – 2018-07-01
  • less drop detect accuracy? (get back some old behavior)
  • inaccurate wave get less weight
  • fix time to hit using intersect instead of my current location

• 0.01e21 – 2018-06-30
  • MeleeRumble ‒ APS: 67.45% (12th), PL: 390-13 (14th), Survival: 37.14%
  • based on 0.01e19
  • use gradle to generate jar automatically
  • use proguard to remove dead code

• 0.01e20 – 2018-06-28
  • use different surf kernel density

• 0.01e19 – 2018-06-28
  • better energy drop tracking
  • detect wave hits on all enemies

• 0.01e18 – 2018-06-25
  • MeleeRumble ‒ APS: 67.53% (12th), PL: 389-14 (15th), Survival: 37.79%
  • rewrite medianScale
  • surf a wave longer based on sourceError

• 0.01e17 – 2018-06-24
  • MeleeRumble ‒ APS: 67.57% (12th), PL: 389-14 (15th), Survival: 38.35%
  • based on 0.01e14
  • migrate fix from 0.01e16

• 0.01e16 – 2018-06-24
  • MeleeRumble ‒ APS: 67.65% (12th), PL: 389-14 (15th), Survival: 38.04%
  • fix ptLineDist used instead of ptSegDist, which disables a bunch of movement options.

• 0.01e15 – 2018-06-23
  • MeleeRumble ‒ APS: 67.47% (12th), PL: 389-14 (15th), Survival: 37.66%
  • rebase MeleeMove onto Unit

• 0.01e14 – 2018-06-20
  • MeleeRumble ‒ APS: 67.59% (12th), PL: 390-13 (14th), Survival: 38.06%
  • Back to 0.01e10
  • Switch some part to Unit
  • Pseudocolor inspired by Neuromancer

• 0.01e12 – 2018-06-20
  • MeleeRumble ‒ APS: 67.43% (12th), PL: 390-12 (13th), Survival: 38.93%
  • tune surfing bandwidth in another direction

• 0.01e11 – 2018-06-19
  • MeleeRumble ‒ APS: 67.63% (12th), PL: 388-14 (14th), Survival: 37.84%
  • tune surfing bandwidth

• 0.01e10 – 2018-06-19
  • MeleeRumble ‒ APS: 67.56% (12th), PL: 388-14 (14th), Survival: 38.15%
  • fix onHitByBullet not correctly tracked
  • fix melee wave source using outdated data
  • introducing Unit, where every bot is treated equally. Turned off completely yet.

• 0.01e9 – 2018-06-16
  • MeleeRumble ‒ APS: 67.6% (12th), PL: 388-14 (14th), Survival: 38.14%
  • rework energy management

• 0.01e8 – 2018-06-14
  • MeleeRumble ‒ APS: 67.44% (12th), PL: 388-14 (15th), Survival: 36.74%
  • more accurate power selection

• 0.01e4 – 2018-06-14
  • MeleeRumble ‒ APS: 67.51% (12th), PL: 388-14 (14th), Survival: 35.63%
  • Iterative power selection.
  • Remarkable survival improvement.

• 0.01e2 – 2018-06-14
  • MeleeRumble ‒ APS: 67.35% (12th), PL: 388-14 (14th), Survival: 35.6%
  • Bullet power selection takes who I can hit into account. No longer power 3 bullet on dying bot!
  • Inaccurate aim power due to bugs.

• 0.01e – 2017-11-06
  • MeleeRumble ‒ APS: 67.62% (11th), PL: 389-12 (13th), Survival: 36.51%
  • Rewrite melee movement, implement Neuromancer style “brute force” true melee surfing (only surfing preloaded HOT now).
  • Merge ScaledBot movement graphics & adapt it to show predicted paths like Neuromancer does, pretty cool now!
  • Well it takes me an hour to write, two nights to tune & fix bugs...

• 0.01d1 – 2017-11-05
  • Thanks Skilgannon for mentioning paper "Fast Polynomial Approximations to Sine and Cosine" in his code ;) I'll try some approximated function as the replacement of StrictMath.
    • Use approximations from "Fast Polynomial Approximations to Sine and Cosine" for sin & cos, 3~5x faster!
    • Use floor for faster normalAbsoluteAngle & normalRelativeAngle, 5.4x faster!
    • Use Chebyshev Approximations (http://metamerist.com/cheby/example38.htm) for atan & atan2, 13.86x faster! (my version uses degree 5 while Skilgannon uses degree 4, not sure the speed difference though)

• 0.01d – 2017-11-02
  • MeleeRumble ‒ APS: 63.94% (25th), PL: 362-39 (39th), Survival: 26.62%
  • Fixed a lot of bugs in PIF gun ;)
  • Implemented melee wave tracking (no gun heat waves yet), and no reaction to that information yet.

• 0.01c – 2017-10-31
  • Simple melee movement that dodges a lot of HOT bullets without surfing ;)
  • Simple PIF gun that selects scans randomly

• 0.01b – 2017-10-29
  • Rewrite ScalarBot’s framework from scratch, so that every single object is either static or pooled. No object creation/deletion between ticks or even rounds!
  • Write a melee radar
  • Migrate ScalarBot’s 1v1 radar into melee radar