Grobots - Tournaments - Top Nine

The Top Nine is a king-of-the-hill tournament: each iteration contains only the best nine robots from the previous tournament, plus one challenger. This concentrates the good sides, and shows how a side performs against good opponents, rather than how well they can eat helpless targets.

The tournaments are run in a standard 10x10 world with ten sides for at least 200 (previously 100) rounds. There are considerable variations in the results, which could mean 100 rounds is not enough. This could also be due to effects sides have on each other - with only ten sides, replacing one can substantially affect the others. So don't take the results too seriously.

Current Top Nine

Side	Author	Date Added
Segregated Eaters 3	Matt Burkholder	12 December 2004
Move Zig! 3	ziggystar	16 June 2006
Weird Men D	Matt Burkholder	30 December 2006
The Lunacy	David/Warren	22 April 2012
Active 9	Warren	22 April 2012
RK Bertha	Warren	7 August 2012
Faithless	fmb	3 June 2013
ISI	fmb	3 June 2013
Mice	Szeety	14 September 2016

History

I started with the top ten sides from a larger tournament: Active Large, Active 3, Commune 6, Productive 5f, Productive Plus 2, Eventually 12, Wiser, World Toad 3, Intoxicated 3g, and Teledont 7. These sides were selected for their performance against a larger, weaker set of opponents. They proved not to be the best sides against each other; four of them were quickly replaced by other sides which had fared worse against a larger field.

Date	Challenger	Removed	Comments
2 April 2003	initial ten sides	Teledont 7 and Eventually 12	Active 3 in first place.
3 April 2003	Business Cycle 4 and Gunner 2	Gunner 2
3 April 2003	Feudal	World Toad 3
3 April 2003	Not Quite Wise 2	Feudal
3 April 2003	Bacteria	Intoxicated 3g	Bacteria is resistant to missiles.
3 April 2003	Vaccine 2	Bacteria	Productive Plus 2 in first place.
3 April 2003	Intoxicated Small	Vaccine 2
4 April 2003	Ants 5	Ants 5	Active 3 was second to last!
18 April 2003	Eventually 12	Intoxicated Small	Ev is back. :( Survival is down to 19%.
30 April 2003	Missile Bacteria	Eventually 12	Missile Bacteria in first place.
26 May 2003	Not Quite Harmless	Active 4	Not Quite Harmless in first place.
29 May 2003	Sucker	Active Large	Updates: Productive Plus 3, Business Cycle 4f, old Not Quite Wise.
27 June 2003	Ring of Fire 2	Sucker 2	Updates: Sucker 2, Wiser 2, Commune 8, Missile Bacteria 2.
28 June 2003	Active 5	Productive Plus 3
28 June 2003	Noffee	Wiser 2
29 June 2003	Gunner 2	Not Quite Wise	Gunner again :(
14 July 2003	Untouchable	Active 5	Overall survival 41%, Active 15%.
14 July 2003	Revenge Killer	Noffee
17 July 2003	Feudal 2	Gunner 3	Updates: Not Quite Harmless 3, Ring of Fire 3, Gunner 3.
12 August 2003	Wise Men	Feudal 3	Updates: Commune 9, Untouchable 2f, Business Cycle 5c, Feudal 3.
16 August 2003	Noffee 2	Business Cycle 5c
29 August 2003	Teledont 8	Missile Bacteria 2
30 August 2003	Productive Plus 3	Teledont 8	How come the Productives are still working?
6 October 2003	Gunner 3	Ring of Fire 3
23 December 2003	Eventually... 13	Productive 5f	Update: Revenge Killer 2.
5 January 2004	Bacteria 4	Noffee 3	Updates: Productive Plus 4, Eventually... 14, Noffee 3, Gunner 4. Productive Plus 4 in first place.
11 January 2004	Commune Plus	Wise Men	Update: Untouchable 3. Commune Plus in first place.
11 January 2004	Frog Celestial 2	Bacteria 4	Untouchable 3 in first place.
22 February 2004	Revenge Garden	Not Quite Harmless 3	Untouchable 3 in first place.
22 February 2004	Wise Men 2	Frog Celestial 2	Untouchable 3 in first place.
21 April 2004	Business Cycle 6	Gunner 4	Untouchable 3 still in first place.
7 August 2004	Wild Men	Revenge Garden	Untouchable 3 still in first place. Survival down to 32%.
11 December 2004	Weird Men	Eventually... 14	Updates: Wise Men 3, Business Cycle 7. Untouchable 3 still in first place (29% to Productive Plus 4's 19%).
12 December 2004	Segregated Eaters	Wise Men 3	Untouchable 3 still in first place.
12 December 2004	Wise Men 2	Wild Men	It looks like Wise Men 3 was a downgrade.
13 November 2005	Grudge 4	Weird Men	Segregated Eaters in first place. Overall survival 28%; error > 2%.
14 November 2005	Shepherds	Grudge 4	Untouchable 3 18%, Shepherds 17%. Overall survival 36%.
16 November 2005	Wise Men 2 (without José)	Wise Men 2	Untouchable 3 in first place.
11 April 2006	Walled City	Shepherds	Updates: Commune 10, Untouchable 4, Commune Plus 2. Segregated Eaters in first place.
3 June 2006	Unproductive	Revenge Killer 2	31% survival.
16 June 2006	Move Zig! 3	Wise Men 2 (without José)	Update: Segregated Eaters 1.1. Move Zig! 3 in first place.
22 July 2006	Shield + Cowards 2f	Untouchable 4	Update: Walled City 2.
30 December 2006	Weird Men C	Productive Plus 4	Updates: Unproductive 2, Segregated Eaters 2.
22 September 2007	Cyclops	Business Cycle 7	32% survival. Move Zig! 3 in first place.
5 January 2008	Revenge Killer 3	Commune 10	Update: Segregated Eaters 3.
11 January 2009	Adaptive 4c	Commune Plus 2	Updates: Revenge Killer 4, Weird Men D.
13 September 2009	Massive Beast 2 Copia	Walled City 2	20% survival.
5 January 2010	The Horde	Unproductive 2
21 April 2012	ARG 1.1	Shield + Cowards 2f
22 April 2012	Active 7	Adaptive 4c	Replaced Massive Beast with The Lunacy, to limit the number of variants. 15% survival.
7 August 2012	Blaster Bertha	Cyclops	Update: Active 9
3 June 2013	Faithless	ARG 1.1	Update: Revenge Killer 5. Replaced Blaster Bertha with RK Bertha.
3 June 2013	ISI	The Horde
14 September 2016	Mice	Revenge Killer 5	Unarmed one-type side!

Survival increased rapidly as bad sides were removed, and has usually been 30-35%.

Some results

If you can't beat them, join them

(7 August 2012) Active deals with the Massive Beast derivatives by dodging their shots. Big Bertha copies their distasteful strategy of seeding with a large long-range fighter, and dramatically improves on Active's performance. I'll miss Cyclops, though.

Rank	Side	Author	Score	Nonsterile survival	Early death	Late death	Early score	Fraction	Kills
1	Blaster Bertha	Warren	45.9%	49%	32%	27%	21.8%	93%	35%
2	Active 9	Warren	27.2%	30%	44%	47%	16.7%	91%	23%
3	The Lunacy	A madman	5.2%	7%	72%	76%	8.4%	78%	7%
4	The Horde	David and Midnight	4.6%	6%	69%	81%	9.4%	77%	7%
5	ARG 1.1	David	4.5%	4%	80%	77%	6.9%	87%	10%
6	Segregated Eaters 3	Matt Burkholder	3.1%	4%	74%	84%	5.5%	66%	5%
7	Revenge Killer 4	Warren	3.0%	4%	67%	87%	8.6%	68%	4%
8	Move Zig!	ziggystar	2.8%	5%	67%	83%	5.8%	51%	3%
9	Weird Men D	Matt Burkholder	2.3%	3%	65%	91%	7.0%	74%	3%
10	Cyclops	Warren	1.3%	5%	58%	89%	9.8%	27%	3%
Overall (1025 rounds):				12%	63%	68%			189%

Finally!

(5 January 2008) Finally, the last of the original sides is gone: no more Commune! Well, no more two versions of Commune, anyway - Commune Plus is still around.

Side	Author	Score	Nonsterile survival	Early death	Late death	Early score	Fraction	Kills
Move Zig! 3	ziggystar	18.3%	38%	32%	43%	9.0%	47%	13%
Revenge Killer 3	Warren	16.0%	34%	21%	57%	15.6%	45%	13%
Cyclops	Warren	13.1%	48%	20%	41%	15.1%	28%	11%
Segregated Eaters 3	Matt Burkholder	11.1%	25%	38%	59%	7.1%	39%	13%
Unproductive 2	Devon	9.2%	32%	28%	55%	11.1%	28%	13%
Shield + Cowards 2f	Alexander (aka nfwu)	9.1%	29%	60%	27%	5.2%	31%	4%
Walled City 2	Warren	8.5%	25%	38%	60%	10.0%	32%	12%
Commune + 2	Devon/Warren	7.1%	18%	49%	66%	8.2%	38%	8%
Weird Men C	Matt Burkholder	5.3%	26%	26%	65%	10.3%	18%	9%
Commune 10	Devon	2.4%	10%	44%	83%	8.5%	23%	5%
Overall:			29%	36%	56%

Weird

(30 December 2006) 300 rounds (why not?). Survival is still falling. Especially Productive Plus' - it suffers greatly from sharing the corners with three other sides.

Side	Author	Score	Nonsterile survival	Early death	Late death	Early score	Fraction	Kills
Segregated Eaters 2	Matt Burkholder	24.9%	50%	20%	37%	10.8%	48%	20%
Move Zig! 3	ziggystar	22.7%	51%	23%	34%	13.0%	44%	15%
Shield + Cowards 2f	Alexander	11.0%	34%	47%	36%	7.4%	32%	6%
Walled City 2	Warren	10.5%	29%	33%	57%	11.9%	35%	13%
Unproductive 2	Devon	8.1%	27%	22%	65%	12.0%	27%	13%
Weird Men C	Matt Burkholder	5.9%	28%	23%	63%	11.2%	20%	10%
Commune + 2	Devon/Warren	5.8%	14%	37%	77%	10.1%	38%	7%
Commune 10	Devon	4.5%	15%	41%	75%	9.8%	27%	6%
Business Cycle 7	Matt Burkholder	3.4%	16%	46%	70%	9.5%	18%	8%
Productive Plus 4	DvF and Devon	3.1%	9%	77%	61%	4.2%	34%	3%
Overall:			27%	37%	57%

Narrowing

(4 June 2006) Notice how much closer the scores have become. This is partly due to an increase in fortification, so Segregated Eaters no longer does so well. I don't understand the drop in survival. 150 rounds, so errors are 2.4%-4.5% - comparable to the differences between scores.

Rank	Side	Author	Score	Survival	Early death	Late death	Early score	Fraction	Kills
1	Productive Plus 4	DvF/Devon	14.2%	29%	61%	26%	7.4%	50%	9.3%
2	Unproductive	Devon	13.3%	45%	31%	34%	9.9%	29%	11.9%
2	Commune Plus 2	Devon/Warren	13.3%	30%	33%	55%	11.3%	42%	9.5%
4	Segregated Eaters	Matt Burkholder	12.0%	36%	29%	49%	9.1%	30%	12.2%
5	Walled City	Warren	10.8%	36%	29%	49%	11.2%	30%	12.6%
6	Wise Men 2 (without José)	Matt Burkholder	8.5%	33%	17%	60%	11.0%	25%	16.4%
7	Commune 10	Devon	8.3%	25%	39%	59%	9.9%	30%	6.9%
8	Untouchable 4	Warren	7.9%	25%	41%	58%	9.1%	30%	6.1%
9	Business Cycle 7	Matt Burkholder	7.5%	29%	44%	48%	9.3%	24%	8.5%
10	Revenge Killer 2	Warren	4.1%	21%	26%	71%	11.8%	18%	6.8%
Overall:				31%	35%	52%

Segregated Eaters' score is not survival * fraction because fraction counts sterile survival, which happens fairly often for this side. The identical death rates for Walled City and Segregated Eaters are a coincidence, not a mistranscription.

History

(11 April 2006 - wow, I haven't updated this for two years) Weakening forcefields has, predictably, hurt all their users. Shepherds has fallen off the top 10. Here are the results in graph form:

This doesn't tell the whole story, because currently graphs count all scores after elimination as zero. So sides like Segregated Eaters, which often wins by elimination, don't look very good in the graph. Also, this is a graph of biomass, not biomass fraction, so sides which get a high fraction by killing rather than growth don't look good. The top line in the graph above is Commune Plus 2, because it grows large. Segregated Eaters took first place by fraction, but is one of the lowest lines on the graph.

Back in business

(21 April 2004) The new version of Business Cycle, in addition to being fun to watch, kicked off Gunner. Yay.

Rank	Side	Author	Score	Survival	Early death	Late death	Early score	Fraction	Efficiency
1	Untouchable 3	Warren	22.1%	64%	8%	30%	14.3%	34%	54%
2	Productive Plus 4	DvF/Devon	21.1%	45%	48%	13%	9.7%	47%	44%
3	Commune Plus	Devon/Warren	13.6%	40%	24%	47%	10.9%	33%	47%
4	Commune 9	Devon	10.1%	38%	23%	51%	10.9%	27%	37%
5	Business Cycle 6	Matt Burkholder	9.0%	56%	19%	31%	13.5%	16%	21%
6	Wise Men 2	Matt Burkholder	8.0%	39%	13%	55%	11.4%	18%	14%
7	Eventually 14	Devon	5.9%	20%	50%	60%	5.9%	28%	25%
8	Revenge Killer 2	Warren	4.7%	27%	29%	62%	10.4%	16%	7%
9	Revenge Garden	Devon	3.1%	16%	36%	75%	9.2%	17%	10%
10	Gunner 4	Warren	2.2%	10%	65%	71%	3.7%	23%	0%
Overall:				36%	31%	48%			31%

I included the Efficiency statistic ((biomass - seed) / income), but it doesn't seem to be useful.

Musical missiles

(23 December 2003) The newly upgraded Eventually scores surprisingly well. Notice Gunner and Productive's tied score, and how close their tiebreaking survival is?

Rank	Side	Author	Score	Survival	Death rates		Early Score	Fraction	Solar
Rank	Side	Author	Score	Survival	Early	Late	Early Score	Fraction	Solar
1	Revenge Killer 2	Warren	29.0%	63%	10%	30%	15.2%	46%	0%
2	Untouchable 2f	Warren	18.2%	66%	8%	28%	13.3%	28%	38%
3	Commune 9	Devon	13.6%	44%	12%	50%	10.5%	31%	55%
4	Not Quite Harmless 3	Warren	9.1%	70%	2%	29%	13.9%	13%	0%
5	Eventually 13	Devon	9.0%	33%	35%	49%	7.9%	27%	100%
6	Wise Men	Matt Burkholder	8.0%	27%	15%	68%	9.6%	30%	8%
7	Productive Plus 3	DvF/Devon	4.2%	13%	58%	69%	6.4%	33%	100%
8	Noffee 2	Warren/Devon	3.7%	33%	7%	65%	10.3%	11%	25%
9	Gunner 3	Warren	2.5%	12%	38%	81%	6.4%	21%	100%
10	Productive 5f	Daniel von Fange	2.5%	11%	57%	74%	6.3%	23%	100%
Overall:				38%	24%	50%			35%

More good fighters

Rank	Side	Author	Score	Survival	Death rates		Early Score	Fraction	Killed
Rank	Side	Author	Score	Survival	Early	Late	Early Score	Fraction	Killed
1	Revenge Killer	Warren	29.2%	61%	5%	36%	15.9%	48%	25.5%
2	Commune 9	Devon	14.3%	45%	23%	42%	10.2%	32%	13.7%
3	Untouchable 2f	Warren	12.9%	48%	14%	44%	11.2%	27%	10.6%
3	Productive 5f	Daniel von Fange	12.9%	29%	33%	57%	10.5%	44%	7.8%
5	Not Quite Harmless 3	Warren	10.0%	52%	6%	45%	12.3%	19%	6.1%
6	Wise Men	Matt Burkholder	9.0%	29%	16%	65%	9.5%	31%	13.5%
7	Missile Bacteria 2	Devon	3.4%	28%	4%	71%	11.3%	12%	5.4%
8	Ring of Fire 3	Devon	2.9%	16%	43%	72%	5.5%	18%	2.2%
9	Business Cycle 5c	Matt Burkholder	2.8%	22%	24%	71%	6.0%	13%	8.2%
10	Feudal 3	Devon	2.6%	20%	29%	72%	7.4%	13%	6.7%
Overall:				35%	20%	56%

Not at all harmless

(26 May 2003) 333 rounds (forgot to set a limit), but who cares?.

Rank	Side	Author	Score	Survival	Death rates		Early Score	Fraction	Kills
Rank	Side	Author	Score	Survival	Early	Late	Early Score	Fraction	Kills
1	Not Quite Harmless	Warren	24.7%	80%	5%	16%	14.5%	31%	9.4%
2	Missile Bacteria	Devon	22.7%	54%	8%	42%	12.8%	42%	15.2%
3	Commune 7	Devon	11.6%	41%	21%	48%	9.9%	28%	10.9%
4	Not Quite Wise 2	Warren	10.3%	52%	16%	38%	13.3%	20%	7.5%
5	Productive Plus 2	Daniel von Fange/Devon	8.7%	21%	57%	51%	7.8%	42%	7.4%
6	Wiser	Warren	8.2%	49%	15%	42%	13.6%	17%	12.1%
7	Productive 5f	Daniel von Fange	5.7%	14%	67%	57%	5.5%	40%	4.6%
8	Business Cycle 4	Matt Burkholder	3.3%	22%	38%	65%	6.0%	16%	7.7%
9	Active Large	Warren	2.4%	9%	30%	88%	8.3%	28%	16.2%
10	Active 4	Warren	2.3%	11%	33%	83%	8.3%	20%	8.9%
Overall:				35%	29%	50%

Surprise!

Rank	Side	Author	Score	Survival	Death rates		Early Score	Fraction	Kills
Rank	Side	Author	Score	Survival	Early	Late	Early Score	Fraction	Kills
1	Missile Bacteria	Devon	31.2%	60%	4%	38%	14.6%	52%	17.6%
2	Commune 7	Devon	13.8%	45%	20%	44%	12.4%	31%	11.6%
3	Productive 5f	Daniel von Fange	11.7%	19%	64%	47%	7.0%	62%	6.3%
4	Business Cycle 4	Matt Burkholder	9.1%	25%	44%	55%	7.2%	37%	9.3%
5	Productive Plus 2	Daniel von Fange/Devon	8.9%	17%	60%	57%	6.7%	52%	7.1%
6	Wiser	Warren	6.6%	38%	25%	49%	13.3%	17%	9.4%
7	Active 3	Warren	5.5%	24%	24%	68%	9.9%	23%	11.6%
8	Not Quite Wise 2	Warren	5.1%	32%	15%	62%	14.9%	16%	6.5%
9	Active Large	Warren	3.9%	11%	26%	85%	9.6%	36%	17.8%
10	Eventually 12	Devon	3.9%	8%	77%	65%	4.4%	49%	2.7%
Overall:				28%	36%	56%

Bacteria getting kicked off

Rank	Side	Author	Score	Survival	Death rates		Early Score	Fraction	Average value:
Rank	Side	Author	Score	Survival	Early	Late	Early Score	Fraction	Killed	Lost
1	Productive Plus 2	Daniel von Fange/Devon	17.8%	26%	56%	41%	8%	68%	23K	26K
2	Active 3	Warren	16.3%	42%	16%	50%	10%	39%	40K	18K
3	Commune 6	Devon	16.2%	35%	29%	51%	9%	46%	22K	12K
4	Active Large	Warren	11.3%	31%	26%	58%	9%	37%	47K	21K
5	Productive 5f	Daniel von Fange	10.0%	18%	64%	50%	6%	55%	12K	22K
6	Not Quite Wise 2	Warren	9.3%	46%	12%	48%	15%	20%	15K	33K
7	Business Cycle 4	Matt Burkholder	7.0%	17%	37%	73%	7%	41%	16K	14K
8	Wiser	Warren	6.2%	36%	13%	59%	13%	17%	16K	29K
9	Vaccine 2	Devon	3.2%	28%	7%	70%	14%	11%	14K	40K
10	Bacteria	Devon	2.6%	18%	20%	77%	8%	15%	13	32K
Overall:				30%	28%	59%			20K	25K

First Round

Rank	Side	Author	Score	Survival	Death rates		Early Score	Fraction	Average value:
Rank	Side	Author	Score	Survival	Early	Late	Early Score	Fraction	Killed	Lost
1	Active 3	Warren	18.7%	32%	34%	52%	13%	59%	20K	16K
2	Active Large	Warren	18.2%	21%	17%	75%	16%	87%	34K	20K
3	Productive Plus 2	Daniel von Fange/Devon	16.6%	20%	67%	39%	9%	83%	13K	18K
4	Productive 5f	Daniel von Fange	12.5%	15%	71%	48%	8%	83%	8K	17K
5	Commune 6	Devon	9.7%	23%	43%	60%	11%	42%	9K	15K
6	Wiser	Warren	9.1%	18%	25%	76%	16%	50%	9K	26K
7	Intoxicated 3g	Daniel von Fange	4.2%	5%	70%	83%	7%	85%	16K	9K
8	World Toad 3	Devon	3.9%	7%	60%	82%	8%	55%	9K	12K
9	Eventually 12	Devon	3.7%	4%	87%	69%	3%	94%	3K	11K
10	Teledont 7	Matt Burkholder	3.3%	6%	61%	85%	7%	55%	9K	12K
Overall:				15%	54%	68%			13K	16K

(April 2, 2003) There are several strange things going on here:

60% of the sides use missiles. Ants' good survival in other tournaments is suddenly looking less surprising. Maybe you should include good missile defense in your next side. :)
All the sides except Commune are fairly agressive, so survival is an amazing 15%. That implies at least half of the rounds were won by elimination.
What works well for slaughtering Bacteria and fending off rampaging algae may not be optimal against serious sides. :)
Many good ideas aren't represented here because there hasn't been a good implementation of them yet.

So don't be too disappointed at the sort of sides included here, or too quick to conclude that violence rules Grobots. Things should improve.

Grobots by Devon Schudy and Warren Schudy