kar: let every group list their favorite parameters and we'll have atleast one
	tournament per configuration.

	m	n	p	q	v
1	50	50	.005	.01
2	100	100	.02	.2	7
3	40	40	.008	.002	2
4
5	40	40	.005	.002
6	75	75	.002	0
7	25	25
8	50	50
9	50	50	.005	0


kar: we'll run multiple tournaments for each configuration to eliminate
	the random factors.
kar: we'll stick to 5000 rounds as the maximum.

kar: we'll include some of last years players.
	Black Plague
	Hunters and Farmers
	Kill Kill Kill

kar: this is the last discussion class.
	any new ideas !!

pv: at the beginning the situation is different, so we reproduce a lot
	and spread in the beginning.

becky: we noticed that all players were spreading a lot in the beginning.

mss: one thing is spreading and the second thing is staying away from friends.
	we tried to combine these effects.

jeff: we did something similar, but by estimating the density of the board.

hari: we reproduce in a spiral motion.

zz: have you thought of anyone else impersonating you guys.

miguel: there's a 2-3 turn handshaking for communication, so its quite
	unlikely.

suv: if you're just sending information about food to everyone, wont all your players cluster around that food ?

miguel: the idea is to transform into a farm if the conditions are good.

pavan: we tried the offline training method to utilize the food density.

kar: how much offline training was needed.

sov: we just ran about 15-20 games but the results were consistent.

kar: if there are no new ideas, lets think about older ideas.
	Can you comment about which ideas blossomed and which ideas
	didnt pan out..

chris: all the last years reports said that estimating p and q was too hard.
	but I think its not that difficult and lot of current groups
	are doing it.

john: we dont really care about p and q, but just the distance to food.

bb: other thing we did was that a newborn player, we make it aggressive.
	but we still dont know whether it was beneficial or not.

rean: we tried some genetic algorithms as mentioned by most groups
	last year. also some adaptive players, but found that the adaptive
	phase doesnt necessarily converge.

bogdan: we had two organisms come together and meet and exchange information
	and the children would be mixed.

kar: any other ideas ?

suv: initially we thought the organisms would start as a farm, but we found
	it impractical in the actual tournaments.

sov: we found that dormancy is OK.

mss: we also keep track of food history.

kar: I'll ask each of you to name a group which'll do well in the tournament.

jeff: I think Group8

deepti: Group7

Ratul: I agree with Deepti

Rob: I would go with 7 too

sov: I would choose group8

sasha: I would say group8 as well. they did well in harsh conditions.

pj: 8

alph: 8. they're really good finding food.

vipul: 8. initially they swarm and then they convert conservative.

andy: 7

james: 7

chris: 7

bc: 1. this is the only group which cares about learning.

suv: 1. because they estimate p and q which is good.

john: 8, for all the above reasons.

rean: 8

bb: g7

akash: 8

becky: 7. I like the grid pattern

pv: 7

miguel: 7

mss: 7. they also do well in harsh conditions.

andrey: 1

peter: 7

anya: 7

hari: 7

eric: 7

kar: lets talk about presentations.

miguel: is pdf okay for the report ?

kar: looking back at the project, last years' students thought there was
	lot of interesting space to be explored.

miguel: I think last year everyone jumped up quickly and tried to estimate
	p and q, but ultimately said it was a bad idea.

kar: dont forget to acknowledge other people's ideas and code if
	you borrowed.

kar: the next three problems are going to be new but that does not mean
	they're going to be more difficult.

mss: we were constantly trying to make the organism adapt to all kind
	of conditions but I would say its much much easier to concentrate
	on one kind of condition.

kar: any other parting thoughts ?

Ratul: maybe next time make them carnivorous.