Changelog

# Changelog

TODO: Do proper wave surfer, at least for 1on1. I.e. be smarter with point
choice, some (weak) attempt done in v6.2

TODO: try to increase melee bullet power, HawkOnFire often get more bullet
damage bonus than me.

TODO: make center of the field more dangerous in melee, so bot moves closer to
the wall. simple.Wall has amaizing survival rate, with simple trick. See
voidious.Diamond too.

TODO: count only passed wave when updating firing stats.

TODO: calculate danger of a point according to angle at which all bots are
seen. The smaller the better, since I see all bots with less radar sweeps.

TODO: check linear and circular guns for the check when last points is not
available at the fire request time, so one had to extrapolate to this time.

TODO: take in account wave fired by other bots to calculate my guns stats

TODO: ramming attack if guns fail

TODO: chose best firing angle gun, the one which hits most targets

TODO: find optimal number of new variants of the path on every tick, while
yet avoid skipped turns.

TODO: increase bullet power in 1on1 to something like 1500/dist. Survival
bonus is small but bullet bonus is increases with bullet energy. So it better
aim for higher bullet damage than survival bonus.

TODO: see why in melee, the bot likes to sit still near the wall or conner, it
start to happen near v1.5. Looks it due to changes introduced in v1.4
May be add maximum staying time in a given point. Looks like survival in melee
is a bit worse for v1.4 than for v1.3

TODO: see why bot does not fire when energy drops below 1. Usually, it does
not fire when remaining energy is 0.6

TODO: take in account guess factor stats collected by other bots

TODO: move enemy firing logic to enemy gunManager

TODO: make guess factor flattener

TODO: make antiAntiGrav gun: which tries to predict enemy motion based on
anti-grav physics

TODO: improve flipLastGuessFactorGun, or make new one with cross correlation
analysis. Looks like there is anti correlation between odd/even waves.
Looks like it was quite crazy idea, though it did help to boost APS in
roborumbles.

TODO: play with fighterBot danger radius. When I  changed 100 -> 200, I saw
drastic drop of APS in melee for my test set. But is seems to keep master bot
from crossfire situations, so may be I need it to depend on number of
opponents.

TODO: work on guess factor lagging gun. There are typically 3-4 wave of my
bots in the air, so when we fire, we do not know last 3-4 guess factors, so we
should look for correlation between 3-4 shots in the past. Quick octave plot
reveals that there is a correlation for the lags up to 4 (sometimes 5) even
for the top bots  and than it is lost. So we should make a map
oldLaggingGF->firingGF, which is 2D distribution map and look for
correlations.

TODO: if radar looks other way, onHitByBullet -> scheduledEdrop wrongly force  enemy to emmit 
non existing bullet. Need to check the last scan time and scheduledEdrop time
to avoid this.

DONE: analyze missed bullets to see which gun the enemy was using. Cannot be
done, this event is not provided for enemy bullets

# v12.8

Logic change

- when look for safety shadow, allow bullet and wave to travel at least 2 tics

Improvements

- use bisection to find bullet intersecting a wave
- hit-miss tree has its own size limit

# v12.7

Bug fixes

- in the bullets shadow code: the shadow size was often overestimated

# v12.6

Logic change

- gunTreePoint: undo increased weight of average lateral speed. Back to v12.3
- gunTreePoint: use last lateral speed change time instead of round number

# v12.5

Logic change

- gunTreePoint: increased weight of average lateral speed
- gunTreePoint: lateral speed average is lateral at every time

# v12.4

Logic change

- using round number (instead of number of bots) as dimension for gunTreePoint

Improvement

- added setRealHitWidth method

# v12.3

Logic change

- Flattener enabling threshold is higher

# v12.2

Logic change

- non learning guns flattener redone

Improvement

- added safety net against supplied zero bot danger width

# v12.1

Logic change

- retune flattener

# v12.0

Logic change

- Use safety zone concept. This is for non learning guns: if a bot was not hit
  in a certain place it is probably safe place. Once certain threshold is
  reached, switch to anti-GF behaviour, where all visited stats are avoided.
- kdtreeGuessFactorGun added gfHit width attribute
- probability density use smallest value as a base
- realHitsGun initializes some values/attributes

Bug-fix

- gfHit coordinates are properly flipped now

# v11.9a

Logic change

- Added inferred gf hit attribute, for gf which I try to deduce from the game
  symmetry
- changed calculation of decay for GF, it takes in account only round
- new function which returns tic Round number

## v11.9

Logic change

- keep track of virtual/real hits/waves
- weights of above events are a bit different
- TODO: there is double counting of virtual bullets if real hit happens

## v11.8

Logic change

- using average lateral velocity instead of average speed for kdTree
  parameters

## v11.7

Logic change

- added danger for slow motion while on the wave

Enhancements

- added profiling toggle

## v11.6

Enhancements

- added calculateDangerFromEnemyWaves for dangerPathPoint
- wave danger using variable for danger not a magic number
- removed duplicate code
- bullet danger checks for a border case and normalizes more than 360 degree angles
  separation

Bug-fix

- wave danger with bot corridor is properly calculated, before it was using
  only statistical GF danger and no bullets danger
- fixed bot shadow function, it was not rotating bot shadow with respect to head on
  angle


## v11.5

Logic change

- anti-rammer improvements attempt
- added special danger for points inside bots overlap zone
- wave danger use the full bot shadow, not just nearest gf danger

Enhancements

- added euclidean distance calculation function

## v11.4

Logic change

- anti-rammer improvements attempt

## v11.3

Logic change

- gunTreePoint weights redone
- added bullet velocity as coordinate for gunTreePoint

## v11.2

Logic change

- gunTreePoint weights redone

Enhancements

- off-line analysis tracks target bot name
- kdTree switched to weighted one, to simplify logic

## v11.1

Logic change

- use kdTreeGun(100) instead of kdTreeGun(10) for flattener

Enhancements
- can log kdtree points for off-line analysis

## v11.0

Enhancements

- list safe removal of fighterBot waves from myWaves list
- bug fix: virtual waves were fired twice, and were not properly removed

## v10.9

Logic change

- fire virtual wave only first 200 tics
- asking 1 mS Thread.Sleep if there is spare time
- gunTreePoint uses averageSpeed
- gunTreePoint uses number of bots back


## v10.8

Logic change

- add virtual bullets to virtual wave only for the first 200 tics

Enhancements

- remove bullets which miss (to update gun hit rates) only for the first 200
  tics. After this time it just give small correction but consumes a lot of CPU.

## v10.7

Logic change
- different kdTreePoint calculation formula

Enhancements
- fixed center of mass formula (still unused)

## v10.6

- profiling a lot of internal functions, speed should be higher now
- added ability to add bullet to a virtual wave (disabled now)

Enhancements
- improved profiler output
- added cpuManager to calculate cpuConstant
- added timer class for future use
- aiming condition does not use time in the future
- gunTreePoint holds its cache the whole round

## v10.5

Enhancement
- remove and count bullets which do not hit target as wave approaches
  - as wave approaches we can deduce which bullets will miss for sure,
    if we count it, we can adjust guns hit probabilities before wave reaches,
    enemy. This should help especially at the beginning of a game, otherwise 1st 3 wave
    are fired with a 1st default gut on long distances.

Bug fix
- energy disabling condition was wrong, so bot disabled itself with last fire

## v10.4

Enhancements
- speed up of gun performance evaluation
- nicer guns report

## v10.3

Logic modification
- master bot fires virtual waves each tic
- relaxing corridor overlap check
- small speed up for 1on1 case

Enhancements
- profiler outputs total exec time
- a bit of speedup of profiler start code

Bug fix
- wrong gun stats report

## v10.2

Skipped

## v10.1

Logic modification
- fixed bugs in bullets' safety corridor logic
- remove safety corridor requirement to be large

Enhancements
- added bulletShieldGun (currently unused)

## v10.1

Logic modification
- less agressive survival

## v10.0

Logic modification
- smart energy management, attempt to survive instead of fire energetic bullets

Enhancements
- keep track of last fired bullet energy
- new functions:
  - energyGainPerTickFromEnemy
  - energyDrainPerTickByEnemy
  - enemyHitRate
  - gunHeat
  - bulletGiveBackByEnergy

## v9.8

Logic modification
- visited stats danger, enables only if unknownGun perfroming better than a threshold

Enhancements
- added new method getBestGunAgainstBot

## v9.7

Logic modification
- enemy wave gf have contribution from real hits and visited gf, and not only real hits

Enhancements
- new helper class to do math with arrays
- getGFdanger returns probability density instead of raw numbers

## v9.6

Logic modification
- if we get hit by bullet, we immediately update firing GF danger in waves in the air
  - hopefully my bot will not be hit by several bullet with the same GF in a row

Enhancements
- more efficient wave creation
- simplified wave creation
- Changed wave creation and update logic
- use a common method to calculate enemy wave danger
- added method to calculate incoming wave danger
- added draft of time decay capabilities for kdTree gun (bad round border crossing)
- record gfHit fired time for analysis
- gfHit keeps track when it was fired
- wave with bullets marks head on direction
- master bot set wave target when fires

## v9.5

- Fixed wave surfer logic. Consider all points on a wave, not just the first one

## v9.4

- enhancement: prevent aligning of the bot along x-axis when it reaches final point

## v9.3

- added waveSurfer motion (only one wave)
- fixed anti-kdTree gun array creation and storage
- show white box where the wave hits the bot path
- graphics shows current escape angle on the incoming waves

## v9.2

- fixed bug in kdTree. kdTree cluster is returned sorted from worst to best!
  - this bug was introduced around v8.4, also cached cluster worked wrong way
- added calcFlipedLateralVelocityPositionFromCoord to utilize game symmetry
- when adding git to the tree, adding a mirrored to latteral velocity hit too.


## v9.1

- analyze bullet hit bullet to gain info which way enemy fires
- clean up: gunManager use of  method addRealHitGF for hit by bullet cases
- waveWithBullets new method getAngleGF

## v9.0

- added cached versions of MEA and gunTreePoint calculations
- isWithinMEA moved to the base gun as isFiringSolutionWithinMEA

## v8.9

- guess factor type guns check if solution within physical MEA
- firing solution can check if it within physical MEA
- increase safety corridor overlap tolerance by just a bit
- Wave draw physical MEA
- speed up: Wave calculates MEA, posMEA, negMEA only once

## v8.8

- Fixed caching bug introduced in 8.6. There were no cache use. Which led to
  longer calculations.
- Added option to bypass the cache for kdTree guns

## v8.7

- added functions to calculate battle field constrained MEA
- kdTree coordinates calculated VERY different now

## v8.6

- kdTree type guns are using cache (expect less skipped turns)
- guns report their hit names with histogram

## v8.5

- fixed double counting of realHits

## v8.4

version bump to match EvBotNG progress

- kdtreeGuessFactorGun weights points by distance
- some code clean up, should not affect performance

in retrospect: 
- no performance change in 1on1 (slight enhancement)
- performance boost in melee +1.2% APS

## v7.9

- kdTree gun is smoothed too now, logically it is the same v7.6 but I hope
  that CPU load is smaller. I am using gfCoverage from gfHit now.

## v7.8

In retrospect: in this version 1on1 APS performance is on pair with v7.6 (just a
bit smaller). Melee APS performance is a bit better than in v7.7,
but survival is 3% worse. Most likely it is the same as v7.7
within errorbars.

- gf guns now smoothed as it was in v7.6, kdTree is still not
- gfHit now has gfCoverage parameter related to the hit bot shadow size,
  currently it is unused, but intended for kdTree smoothing

## v7.7

In retrospect: in this version 1on1 performance dropped, melee improved (i
think because of the reduced cpu load of kdTree search)

- disabled hit GF spill to nearby bins, i.e. no smoothing

## v7.6

In retrospect v7.6 is so far the best of mine in 1on1 with APS 78.7.

- Wave now have danger from real hits as GF danger values. This replaces old
  virtual GF hits danger.
- Consequently, GF danger from real hits is larger than virtual bullets.

## v7.5

- Improved search of final destination: points outside of battlefield use to
  be thrown and that counted as an attempt. Now such displacement rotated into
  battlefield. So attempts does real search.

## v7.4

- Added ability to track at which GF enemy hits my bot, consequently I can avoid
  such GF even if the gun is unknown

## v7.3

- find a bug introduced somewhere around v7.0: the latteral velocity sign was
  stripped off in the gun kdTree coordinates calculations.

- note: APS wise v7.2 is about the same as v7.1

## v7.2

- retuned GF and Bullets weights for wave danger. GF has a bit higher weights.


## v7.1

- retuned GF and Bullets weights for wave danger. GF has much smaller weights.
  - looks like it is worse against rambots but otherwise higher scores

## v7.0

- Calculations speed up: wave danger calculated based on GF arrays which calculated only once
- wave danger logic redone, probably need to retune coefficients between
  bullets and GF dangers in a wave
- regression: removed lateral velocity logic for GF calculation.
  - I need it back as well as to check logic
- improved readability of profiler output


## v6.9

- redone wave danger weights, now it takes in account combined game GFs danger
- redone probability density calculation for an array with 0 sum

## v6.8

- when normailizing enemy firing solutions, check against unknownGun
performance. If unknownGun has high stats, we have no clue about enemy guns
and should choose some other avoidance strategy. I.e. GF flattener, this part
to be done.
- some extra helper functions


## v6.7

- added ability to do anti guess factor kdTree guns
- added helper class to do array statistics
- increased corners danger radius in 1on1 situations
- if gun is hot use simple head on gun to point the gun in enemy direction

## v6.6

Improvement: Precise bullet intersection of bot body for bullet hit check.
kdTree gun now takes arguments,
kdTree uses advancing (to/from bot) velocity


## v6.5

Skipping v6.3 and v6.4. Only EvBotNG was changing but framework was not.
This synchronizes framework and EvBotNG versions.

- added nice textual plots/histograms for arrays. Very handy for debugging.
- reworked logic of kdTree nodes calculations
- guess factors are smoother now

## v6.2

- in 1 on 1 situation surf only the oldest (hopefully closest wave), this
  seems to help against GF counting guns
- 1on1 point to go calculation favors orthogonal to target motion
- border uncertainty detection is almost 0 now, instead of 1.1
- Added proper circular acceleration gun
- fixed rotation rate calculation
- added MEA marker on enemy wave

## v6.1

Disable fancy bullet energy search. I.e. back to v5.5 bullet logic.

## v6.0

Added track of my score (all game types) and enemy (for 1on1 only). Using it,
attempting to optimize the APS by changing "ideal bullet energy".

## v5.5

Revert bullet energy back for the case of melee, 1on1 stays the same as in v5.4.

## v5.4

Some simplification of the KdTree coordinate code. It is more general and
located in one place now.

Modified bulletEnergy logic to pick 1.95 unless enemy is closer than 140.
This energy selection was spotted in Chase's cs.Nene 1.05 bot.
Test shows that it drastically improve score against stronger bots.


## v5.3

master bot takes gun performance from the KdTree, i.e., it does not use
overall gun performance but it is more adaptive. Rationale, some gun a better
at certain distances/situations.


## v5.2

bug fix: distance to the wall ahead was not assigned for enemy bots stat points

-  botStatPoint now contains timeSinceVelocityChange
-  KdTree has new dimention: timeSinceVelocityChange

## v5.1

v5.0 was better in roborumble but I lost in meleerumble. So some
readjustments:

-  Reduced number of neighbors in kdtreeGuessFactorGun 1000->100.
-  kdtree has new dimentions distance to wall ahead and enemies number.


## v5.0

Added KdTree gun with distance, bulletSpeed, lateral speed, and acceleration
coordinates.

## v4.9

Slightly decreased gun performance weight.

## v4.8

Do not lock radar if no firing solutions are present.

Disable 
-   decrease fire to minimum if we are below enemy (in 1on1 situations),
from the  previous version. This gave significant drop in 1on1 APS.

Redone logic of do not fire. Essentially, I try to decrease bullet power ( to 0
if needed) to keep my energy above enemy if I am on 1on1 or no bullets in the air.

## v4.7

Energy management:
-   increase fire power if we have significant surplus of energy, 
-   decrease fire to minimum if we are below enemy (in 1on1 situations),
-   do not fire into non firing enemy, if it brings my power below the enemy

## v4.6

Weight due to gun performance redone/increased.

Redone the logic for the best firing solution based on weight distribution.

## v4.5

Added aim at everyone logic. 

Commented out the fired count weight when deciding on a robot weight. 

Bot weights calculated only once per tic.

Added debugging graphics: firing angles and gun performance indication.

## v4.4

Fixed gun delayed info accounting.

All gun correct the firing solution to make firing angle along the wall
if the firing bot is sufficiently close to the wall. This should waste less
bullets.

## v4.3

Increased bullet energy a bit, v4.2 change greatly increased survival in
melee, but APS dropped by about 1%.

## v4.2

x.x5 exploit redone to round to smaller side

Modified bullet energy vs distance formula to die off quickly, it improves
survival

## v4.1

Added ability to calculate overall danger of a wave from GF stats, it is not
used for now, since I do not see score improvements. But should help with bot
GF flattener in the future.

Attempt to use x.x5 power bullet detection bug in the bots based on
BasicSurfer.

## v4.0

Bullets and wave danger is calculated on the shadow and guess factors, i.e. no
real space distance calculations. This gives tremendous speed up. This new way
of calculations justifies bumping up the major version.

Safety corridors are properly taken in account for wave danger calculations.

Nice plot of the GF danger for enemy bots wave.

## v3.7

GF updates touch the range of angles covered by a bot body

Added safety corridors capabilities. I.e., bullet and bots shadows.

## v3.6

Each GF gun has its own threshold for goodness decision. As result decaying GF
guns fire now.

## v3.5

Retune the point when GF guns provide results.

## v3.4

New gun: assistedGFGun. This is the guess factor gun with segmentation on
guess factor suggested by the accelerating circular gun at the time of fire.
The idea that there is some correlations, and for some bots I see it. Mine is
certainly have some correlation on such segmentation.

Prepared to collect more statistics at the time of fire, so far it is only
distance.

## v3.3

It is bad idea to be in multiple crossfires, but this is what happens in 1 vs
4 if my bot is in the middle, it become the closest to all, thus everyone fires
at us.  So, In melee when number of enemies = 4, the center of the battlefield has danger.


## v3.2

Added the circular gun which handles an accelerating bot.

## v3.1

Put a safety net around a potential problem spot. When a fictitious wave is
asked to be added from a bit which rams with me before a radar had a chance to
see it. This use to throw the "null pointer exception" which propagates for
the rest of the match.


## v3.0

Remove attractive potential from energetic but non firing bot.

## v2.9

CoreBot caches time, for getTime() calls.

If bot does not fire, it is taken in account as its reduced danger. So my bot
will come closer, if bot has no energy to fire it will ram it.

## v2.8

Reverted target selection logic to v2.6

If a point is the closest to enemy, this enemy danger radius increased, to push
away my bot from being the closest to the enemy. Should increase
survivability.

Even not the closest enemy fires at master bot sometimes, this helps with no
motion freeze when all other bots are the closest to someone else. But some
no so advance bot fire at any target in a view (walls, crazy, etc ).


## v2.7

Redone target selection logic

## v2.6

Revert back number largest number of bots when I fire a wave to everyone: 10->4

## v2.5

When counting hits and missed fire, do it only for waves which passed over a
bot. This avoid counting as missed the waves which are still in the air when
enemy dies, also for the case of virtual bullets, it does not count as a miss
the case when radar was looking away during the wave pass over the bot.

Increase number of bots to which I fire virtual waves at melee: 4 -> 10

## v2.4

Fixed bug, with double counting of the master bot gun shots, which was
introduced in v2.3

## v2.3

Proper counting stats for real enemy hits.

Take in account enemy hit stats, when calculate firing solution danger.

Circular gun use a more general method to find intercept, 
it also tries to project target motion if stats are delayed.

Master bot guns are set in config which can handle different game types.
Enemy bot guns are set in config, as well.

Enemy gunManager is used to provide virtual bullets against master bot

firingSolution plots bullet quality of solution

Added safetyCorridorGun gun to push my bot to safety or random location

## v2.2


In melee radar lock now 2 clicks earlier, than before. Seems to boost score
with my test set.

Circular gun falls back to linear gun scenario, if not enough history
available.

Master bot does not fire linear gun anymore, circular one does its function.


## v2.1

Remove guess factor based guns from the enemy guns array.
These guns were introduced since v1.6 but apparently make the master bot motion more
predictive.  It hearts IWillFireNoBullets roborumble stats,
up to version 2.0 the APS is lower.
Buy now enemy firing and master bot motion logic is the same as it was in
v1.6

Added enemy wave count, to implement lagging guess factor gun.

## v2.0

bug fix: guess factor use the lateral velocity from wrong time

New gun: flipLastGuessFactorGun. I note that many bots including my own
have next wave guess factor close to opposite of the previous wave. This gun
tries to exploit it.

## v1.9
Disabling mostly orthogonal to the enemy direction path search. Looks like 
it drops rating IWillFireNoBullet by about 2%.

## v1.8

Fixed bug introduced in v1.7:  in 1on1 my bot locked itself in a corner, due
to no danger from corner, and only shallow angle search within orthogonal direction to
the enemy.

## v1.7

Enemy bots fire guess factor gun too.

Master bot has new decaying guess factor gun.
Enemy bots also fire decayingGuessFactorGun gun at master bot.

New exact path choice in 1on1 type, try to chose points orthogonal to line to
the enemy.

## v1.6

Radar locks on target when gun is cold enough.
Gun aiming is not executed when gun is too hot.
Added guess factor gun for the master bot gun array. Apparently, in the old
EbBot the guess factor gun did not take in account the lateral velocity
direction. So this realization should outperform it.

Optimization:do not calculate path danger if it is already greater than a
known path. This should reduce CPU usage.

## v1.5

Added bullet shielders detection, and fire with offset to circumvent it.
If bot hit enemy bullet, then enemy wave which carries this bullet is removed,
since it has no danger.

Exact path is attempting to find a better path every turn, but with smaller
number of trials than full search.

All very, strange. Melee score went up, but 1on1 down.
Apparently, increasing maximalPathLength to 50 from 32. makes 1on1 motion more
predictable. Will reduce it to see the difference. Checked, looks like it does not
change anything. Maybe it because path was recalculated every tick.

Looking closely at some battles for the v1.3, I conclude that there were a
glitch in the rumble. When run locally with gui, I do not get stellar 100%
survival against bots which rumble reported as such. For example against
zyx.micro.Ant 1.1. The scores also not so awful for the v1.4 as in the rumble.


## v1.4
Attempt to reduce gun calculations:
Enemy wave directed at us only if master bot is the closest or number of
enemies >=3. Consequently enemy waves with no dangerous bullet are
disregarded.

Exact path is calculated for longer time period and is not updated every tick
any more (unless there is a fire detected or minimal path length is reached).

Added Random gun, to the guns array.

## v1.3
Fixed circularGun bug which was triggered in melee, when no consequent points
are available to figure out velocity rotation.

## v1.2
Added circularGun, enemy fires it too

## v1.1

Fixed linear gun: does not offer future positions outside the field
Fixed contribution of enemy energy to the target weight, it use to select 
most energetic enemies.

Redone weight calculations for best target selection, among other things it
takes in account enemy hit rate and fire count only after some amount of
rounds (5) played.

Bot position danger includes its energy.

## v1.0
v0.7 - 0.9 are skipped since v1.0 has quite drastic changes.

Enemy fires a wave to everyone (might be too CPU intensive).

Do not fire enemy wave if its energy drop is due to the master bot hit.

Count guns success stats.

Chose gun according to its virtual hit probability.

Chose target according to distance, energy, hit probability, survival stats, etc.

Fix a bug: did not detect a bullet with energy very close to 0.1, Diamond
likes to fire such one close to the end of melee.

## v0.6
Improved enemy firing solution generation. Take in account that at fire time
they now target bot stats for the previous tick time.

My bot does not anymore slam the brake, when all enemy are dead. Since bullets
are still flying.

Keep track of bots hit rate and hit by others rate.

Longer default danger path prediction (8 -> 32) should help against rammers



## v0.5
Master bot use position in future to calculate its firing position, thus
correct path prediction of my own bullets. So far I still use only headOn and
linear guns.


## v0.4 - 0.1
lost in history since the frame work was taken from IWillFireNoBullet robot.
So see its version history.