Difference between revisions of "Griezel/Version History"

From Robowiki
Jump to navigation Jump to search
m (typo)
(0.6.3 stats, next step is unclear)
 
(17 intermediate revisions by the same user not shown)
Line 6: Line 6:
 
----
 
----
  
* '''0.4''' -  20060526
+
'''0.6.3''' -  20091019
** Rank      :14  --  APS 64.00  Survival 86.81
+
'''{{RumbleStatsDefault|link=http://darkcanuck.net/rumble/RatingsDetails?game=meleerumble&name=gh.Griezel+0.6.3|rumble=MeleeRumble|scorelabel=APS|score=63.9|rank=17th|win=264|loss=16|plrank=14th|glicko2=1554.2|score2label=Survival|score2=85.94}}'''
** Rank (old) :10  --  Rating 1698
+
* revert to 0.4. (remove all fancy risk-stuff) Use bulletpower enhancement of 0.5.1
** Gun does shoot less into walls
+
* Now I am changing things without any clue anymore
** fixed occasionally lost lock
+
'''0.6.2''' -  20091014
** only switch to new movepoint if it is significantly better (instead of just better)
+
'''{{RumbleStatsDefault|link=http://darkcanuck.net/rumble/RatingsDetails?game=meleerumble&name=gh.Griezel+0.6.2|rumble=MeleeRumble|scorelabel=APS|score=64.19|rank=15th|win=261|loss=19|plrank=21st|glicko2=1560.7|score2label=Survival|score2=85.84}}'''
*** Comment: still holding, despite increased competition
+
* Add back risk at corners and center, no extra risk for current target.
* '''0.3''' - 20060119:
+
'''0.6.1''' -  20091013
** Rank : 10  --  Rating: 1698 (down to 14 - 1689)
+
'''{{RumbleStatsDefault|link=http://darkcanuck.net/rumble/RatingsDetails?game=meleerumble&name=gh.Griezel+0.6.1|rumble=MeleeRumble|scorelabel=APS|score=64.31|rank=14th|win=264|loss=16|plrank=16th|glicko2=1565.3|score2label=Survival|score2=86.77}}'''
** Added Team-support, take distance and perpendicularity as key risksfactors
+
* Higher award for being not the closest.
*** Comment: expected rank 20, got up to rank 10! I am curious if it can hold this place
+
'''0.6''' -  20091011
* '''0.2''' - 20051224
+
'''{{RumbleStatsDefault|link=http://darkcanuck.net/rumble/RatingsDetails?game=meleerumble&name=gh.Griezel+0.6|rumble=MeleeRumble|scorelabel=APS|score=64.27|rank=14th|win=262|loss=18|plrank=16th|glicko2=1566.8|score2label=Survival|score2=86.37}}'''
** Rank : 29  --  Rating: 1663
+
* Try not to be the closest bot to anyone, especially the dangerous one.
** Some minor imporvements outside risk-function
+
** Comment: looks like the only improvement since 0.4 that works is: higher bulletpower at close-to-medium range
*** Comment: minor improvements give minor result
+
'''0.5.4''' -  20091005
* '''0.1''' - 20051219
+
'''{{RumbleStatsDefault|link=http://darkcanuck.net/rumble/RatingsDetails?game=meleerumble&name=gh.Griezel+0.5.4|rumble=MeleeRumble|scorelabel=APS|score=64.01|rank=16th|win=267|loss=16|plrank=14th|glicko2=1567.9|score2label=Survival|score2=85.37}}'''
** Rank : 40  --  Rating: 1658
+
* No risk at corners, some risk in centerfield
** Framework of Gruwel with a simple distance-driven minimumrisk movement
+
** Comment: now the corner/center risk is complete foggy . . .
*** Comment: a bit less than expected, that risk-function really needs some more work
+
'''0.5.3''' -  20091004
 +
* Rank    : 16  --  APS 64.21  Survival 86.50
 +
* Add back some risk at corners
 +
** Comment: where to put what risk becomes a bit blurry
 +
'''0.5.2''' -  20091002
 +
'''{{RumbleStatsDefault|link=http://darkcanuck.net/rumble/RatingsDetails?game=meleerumble&name=gh.Griezel+0.5.2|rumble=MeleeRumble|scorelabel=APS|score=64.36|rank=13th|win=265|loss=15|plrank=14th|glicko2=1569.5|score2label=Survival|score2=86.28}}'''
 +
* Rank    : 13  --  APS 64.52  Survival 86.63
 +
* No risk at all for centerfield and corners
 +
** Comment: performs at par without one of the ''keyfactors'' of Minumum Risk movement
 +
'''0.5.1''' -  20091001
 +
* Rank    : 13  --  APS 64.49  Survival 86.49
 +
* Don't be affraid of the centerfield
 +
* fire with more power at close-to-medium range
 +
** Comment: closing the gap to #10, the max reachable for this bot
 +
'''0.5''' -  20090930
 +
* Rank    : 17  --  APS 63.94  Survival 86.44
 +
* Take into account good enemies (hit me a lot) and weak enemies (I hit them a lot) in riskcalculation
 +
** Comment: surprisingly hardly any difference, only better PL
 +
'''GriezelHawk 1.0''' -  20090830
 +
* Rank : 11  --  APS 64.63 Survival 85.98
 +
* One-time experiment with the movement of HawkOnFire
 +
** Comment: movement only .6 points behind HawkOnFire, improvement should not be that hard.
 +
** Comment: with CT-gun (and survivalistic energymanagement) top 10 still possible
 +
'''0.4''' -  20060526
 +
* Rank      :14  --  APS 64.00  Survival 86.81 (down to 17)
 +
* Rank (old) :10  --  Rating 1698
 +
* Gun does shoot less into walls
 +
* fixed occasionally lost lock
 +
* only switch to new movepoint if it is significantly better (instead of just better)
 +
** Comment: still holding, despite increased competition
 +
'''0.3''' - 20060119:
 +
* Rank : 10  --  Rating: 1698 (down to 14 - 1689)
 +
* Added Team-support, take distance and perpendicularity as key risksfactors
 +
** Comment: expected rank 20, got up to rank 10! I am curious if it can hold this place
 +
'''0.2''' - 20051224
 +
* Rank : 29  --  Rating: 1663
 +
* Some minor imporvements outside risk-function
 +
** Comment: minor improvements give minor result
 +
'''0.1''' - 20051219
 +
* Rank : 40  --  Rating: 1658
 +
* Framework of Gruwel with a simple distance-driven minimumrisk movement
 +
** Comment: a bit less than expected, that risk-function really needs some more work
  
 
mini.Griezel
 
mini.Griezel
* '''0.1''' -  20060410
+
'''0.2''' -  20090902
** Mega rank :21  --  APS 62.83  Survival 84.64
+
* Mega rank :19  --  APS 63.79  Survival 86.55
** Mini rank : 4  --  APS 65.58  Survival 90.39
+
* Mini rank : '''2'''  --  APS 66.65  Survival 92.36
** Griezel v0.3 without teamsupport and with simplified targetselection
+
* Shoot less into walls
 +
* only swith to new movepoint if it is significantly better
 +
'''0.1''' -  20060410
 +
* Mega rank :21  --  APS 62.83  Survival 84.64
 +
* Mini rank : 4  --  APS 65.58  Survival 90.39
 +
* Griezel v0.3 without teamsupport and with simplified targetselection
  
 
[[Category:Bot Version Histories]]
 
[[Category:Bot Version Histories]]

Latest revision as of 22:34, 24 October 2009

Griezel Sub-pages:
GriezelVersion History

0.6.3 - 20091019 MeleeRumble ‒ APS: 63.9% (17th), PL: 264-16 (14th), Survival: 85.94%

  • revert to 0.4. (remove all fancy risk-stuff) Use bulletpower enhancement of 0.5.1
  • Now I am changing things without any clue anymore

0.6.2 - 20091014 MeleeRumble ‒ APS: 64.19% (15th), PL: 261-19 (21st), Survival: 85.84%

  • Add back risk at corners and center, no extra risk for current target.

0.6.1 - 20091013 MeleeRumble ‒ APS: 64.31% (14th), PL: 264-16 (16th), Survival: 86.77%

  • Higher award for being not the closest.

0.6 - 20091011 MeleeRumble ‒ APS: 64.27% (14th), PL: 262-18 (16th), Survival: 86.37%

  • Try not to be the closest bot to anyone, especially the dangerous one.
    • Comment: looks like the only improvement since 0.4 that works is: higher bulletpower at close-to-medium range

0.5.4 - 20091005 MeleeRumble ‒ APS: 64.01% (16th), PL: 267-16 (14th), Survival: 85.37%

  • No risk at corners, some risk in centerfield
    • Comment: now the corner/center risk is complete foggy . . .

0.5.3 - 20091004

  • Rank : 16 -- APS 64.21 Survival 86.50
  • Add back some risk at corners
    • Comment: where to put what risk becomes a bit blurry

0.5.2 - 20091002 MeleeRumble ‒ APS: 64.36% (13th), PL: 265-15 (14th), Survival: 86.28%

  • Rank : 13 -- APS 64.52 Survival 86.63
  • No risk at all for centerfield and corners
    • Comment: performs at par without one of the keyfactors of Minumum Risk movement

0.5.1 - 20091001

  • Rank : 13 -- APS 64.49 Survival 86.49
  • Don't be affraid of the centerfield
  • fire with more power at close-to-medium range
    • Comment: closing the gap to #10, the max reachable for this bot

0.5 - 20090930

  • Rank : 17 -- APS 63.94 Survival 86.44
  • Take into account good enemies (hit me a lot) and weak enemies (I hit them a lot) in riskcalculation
    • Comment: surprisingly hardly any difference, only better PL

GriezelHawk 1.0 - 20090830

  • Rank : 11 -- APS 64.63 Survival 85.98
  • One-time experiment with the movement of HawkOnFire
    • Comment: movement only .6 points behind HawkOnFire, improvement should not be that hard.
    • Comment: with CT-gun (and survivalistic energymanagement) top 10 still possible

0.4 - 20060526

  • Rank :14 -- APS 64.00 Survival 86.81 (down to 17)
  • Rank (old) :10 -- Rating 1698
  • Gun does shoot less into walls
  • fixed occasionally lost lock
  • only switch to new movepoint if it is significantly better (instead of just better)
    • Comment: still holding, despite increased competition

0.3 - 20060119:

  • Rank : 10 -- Rating: 1698 (down to 14 - 1689)
  • Added Team-support, take distance and perpendicularity as key risksfactors
    • Comment: expected rank 20, got up to rank 10! I am curious if it can hold this place

0.2 - 20051224

  • Rank : 29 -- Rating: 1663
  • Some minor imporvements outside risk-function
    • Comment: minor improvements give minor result

0.1 - 20051219

  • Rank : 40 -- Rating: 1658
  • Framework of Gruwel with a simple distance-driven minimumrisk movement
    • Comment: a bit less than expected, that risk-function really needs some more work

mini.Griezel 0.2 - 20090902

  • Mega rank :19 -- APS 63.79 Survival 86.55
  • Mini rank : 2 -- APS 66.65 Survival 92.36
  • Shoot less into walls
  • only swith to new movepoint if it is significantly better

0.1 - 20060410

  • Mega rank :21 -- APS 62.83 Survival 84.64
  • Mini rank : 4 -- APS 65.58 Survival 90.39
  • Griezel v0.3 without teamsupport and with simplified targetselection