Recommend
4 
 Thumb up
 Hide
38 Posts
Prev «  1 , 2  | 

Gaia Project» Forums » General

Subject: Podcast episode on Gaia Project with in-depth gameplay analysis and Automa development history rss

Your Tags: Add tags
Popular Tags: [View All]
James Smith
msg tools
Is there a summary for English speakers?

Was the Automa is designed to replicate the typical behaviours of an experienced human player, eg. in round 3, a human player would have a 50% chance of taking 1 power action, 25% chance of taking two power actions etc.

Or, if the round bonus tile is 2VP for each mine built, a human player would typically score 4VP from this bonus etc.

Is that basically how it was designed? By analysing the behaviour of human players from round to round from a dataset of plays, and then modelling such behaviour probabilistically using the Automa deck?

Or was it more a top-down design where Morten drawing on his experience of playing GP felt the Automa should do more mine building at the beginning of the game, then more tech track advancing in later rounds etc, and then tweaked this during play testing until it felt right?
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Timmi T.
Germany
Brandenburg
flag msg tools
Lass mich!
badge
Geh weg!
Avatar
mbmbmbmbmb
Hatte während einer längeren Autofahrt endlich mal Gelegenheit das mit meiner Freundin anzuhören. Hat Spaß gemacht, danke.
3 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Jonas Egel
Germany
Solingen
Deutschland
flag msg tools
Sehr geiler Podcast, danke! meeplemeeplemeeple

Hat mich gleich aus zwei Gründen brennend interessiert:

1. GP isz mein absolutes Lieblingsspiel! cool

2. Ich versuche derzeit verzweifelt eine Automa-Variante für mein eigenes Brettspiel zu entwickeln und es will mir einfach nicht vernünftig gelingen. Durch den Podcast habe ich dazu neue Anregungen erhalten. arrrh
3 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Jonas Egel
Germany
Solingen
Deutschland
flag msg tools
vanquish7 wrote:
Is there a summary for English speakers?

Was the Automa is designed to replicate the typical behaviours of an experienced human player, eg. in round 3, a human player would have a 50% chance of taking 1 power action, 25% chance of taking two power actions etc.

Or, if the round bonus tile is 2VP for each mine built, a human player would typically score 4VP from this bonus etc.

Is that basically how it was designed? By analysing the behaviour of human players from round to round from a dataset of plays, and then modelling such behaviour probabilistically using the Automa deck?

Or was it more a top-down design where Morten drawing on his experience of playing GP felt the Automa should do more mine building at the beginning of the game, then more tech track advancing in later rounds etc, and then tweaked this during play testing until it felt right?


Today Im too tired to make an english summary for you but in the next days I promise Ill do so.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Morten Monrad Pedersen
Denmark
flag msg tools
designer
badge
Avatar
mbmbmbmbmb
yonek wrote:
vanquish7 wrote:
Is there a summary for English speakers?

Was the Automa is designed to replicate the typical behaviours of an experienced human player, eg. in round 3, a human player would have a 50% chance of taking 1 power action, 25% chance of taking two power actions etc.

Or, if the round bonus tile is 2VP for each mine built, a human player would typically score 4VP from this bonus etc.

Is that basically how it was designed? By analysing the behaviour of human players from round to round from a dataset of plays, and then modelling such behaviour probabilistically using the Automa deck?

Or was it more a top-down design where Morten drawing on his experience of playing GP felt the Automa should do more mine building at the beginning of the game, then more tech track advancing in later rounds etc, and then tweaked this during play testing until it felt right?


Today Im too tired to make an english summary for you but in the next days I promise Ill do so.


Luckily I can read German reasonably well .
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Jonas Egel
Germany
Solingen
Deutschland
flag msg tools
Well i promised it. Here is my translation of the podcast. This summary doesnt contain everything said and is not completely chronological. I rearanged it into the different topics to make it easier to read.

1.) sebastian (moderator): not very experienced gp player but likes the game and is very much interessted in it.
2.) jochen (gaia project fan, very good player) 225 played games 75% of them solo vs automa (ultoma lvl). loves gp and sais its the best game to play with 2 player out there. his favourite faction are the xenos although he doesnt think they are neccessarily one of the stronger factions.
3.) lines (team automa factory) about 40 played games, loves sologaming but not so much eurogames. but he does like the design of gp a lot.

comparison gp and terramystica:
jochen doesnt like when everybody compares gp with terramystica or asks if one of the games needs the other. he never bought a copy of terramystica (while owning more than 100 eurogames).
they agree that gp is the better balanced game.
jochen tells that he asked helge ostertag if there will be an expansion for gp. It is not planned and doesnt look that there will be one.

soloplayer with the automa:
automa was developed since the early beta of gp by three persons. it took about 6 month. it was designed as a 2player game since beginning. later came the virtuell 3rd player for the endgamescoring.
lines explaines that automa is no "a.i" or virtual "player" but more like a "board state simulator".
automa just cares about the aspects of game where there is player interaction. so he doesnt make real moves but he simulates how the board looks if there was another player. the hardest part in developing automa is to make him easy-to-handle, because he shouldnt disturb the players to much in playing, while have him simulate a player very well.
the idea with the splitted automacards came when first there was the idea of two carddecks for much variaty. but in the end they made it the way it is that you only need one deck.
they made automas for the 7 races they did because they asked helge and jens wich where the beguinners races and they didnt want to make all factions because they wanted to have a well balanced and good tested automa so they thought with 14 factions there had been not enough testing data for each race.
jochen tells that it is difficult to win taklons-automa with ivits?! (is that the english term for the faction "der schwarm"?).
sebastian does like the automa-mechanic with the 3 cards that decide if automa ends his turn. lines does point out that those cards do help hindering human players with a good memory to predict what automas last actions per turn are.
lines tells that he and the other two developers does make a good team because they are working very different and do compliment eachothers. Morten does play one or two turns than stop it and develop the next other idea. He (Lines) does like to play games with automa to the end to see who will win in the end and David does focus on certain aspects of the game (f.e. minesbuilding) and does optimize that.

gp balancing:
everyone agrees that gp is very good balanced. jochen sais that every faction (with the one exeption of lantids) is playable at a high level and the strongest faction (the ivits) are not overpowered. He thinks to beat the ivits you should settle very close to him and try to surround him.
jochen doesnt understand how taklons were considered as a beguinners faction. yes they are easy to understand but not easy to play well with. he sais that you can recognice a bad taklonsplayer when he has the brainstone in bowl 3 at the end of the round.
also the automa is well balanced. the highest difficulty "albtrauma" (dont know the english term) is made for the very best player and does score way more than 200 vps.
sebastian asks why automa scores vp for buildings in alliances by counting all his buildings minus 1 as that is a weird number.
Lines explaines that it has been all buildings and the minus 1 comes from playtesting and statistics and shows how important the aspect of automas balance is for the developers.

jochens gp strategie advices:
standart openings are either first-turn-academy or first-turn-mines (2-4 mines). he doesnt recomend first turn researchlab what many beguinners do. The first-turn planetary is only good for few factions (neevla, terrans and sometimes geoden) because if you get no techtile in turn 1 you get no second research step wich is way better then your planeterybonus.
to get the 11 ore for a firstturn-academy you can go 1 step in terraforming but its better when you can do the power-action where you get 2 ore. (thats why ivits is good because they can do this power-action very good and they have no problem with burning power because they dont need power-tokens for building satelites.)
he doesnt take factions if their mechanics compete with those of other players f.e. he doenst like to take taklons if opponent has neevla.
when he chooses faction he looks mostly at the basistechtiles. example: if the +4power action isnt available in a good way he wouldnt choose taklons.
he often ignores the roundbonuses in the first 3 rounds because its much more important to get the economyflow going in the first rounds. when you plan the first round you should plan ahad the second and third round.
he thinks good starting researchtrees are mostly the economic, knowledge and navigation.
he thinks the last step of all research trees are about equal strong. only the 4 qic are little worse than the others.
sometimes it is good to consider making actions in the next round if that helps you getting an important roundbooster or the initiative.
for victorypoints he thinks that in most times the research steps and the advanced-techtiles are more important than the two final scoring tiles.
he thinks one should make about 3 federations. most federations you will ever have is 5-6 and only with the ambas.
he does think playing and analyzing a lot of games helps to get good. automa is a good possibility to play many games and learn from it for games against human players. but automa plays different than good human. (example: automa often takes the roundbooster that his opponent took last, just to annoy)
he does explain much more but its to difficult to remember and translate everything he said here.

other:
they dont understand why the only faction that has an english faction name in the german original-game (the "mad androids") was translated into another (fantasy-)name in the english version (bescods).
they are talking about people dont liking the choice of plastic material for gamecomponents but they agree that they do like the plastic in gp and do think it is good qualitiy (heavy) plastic and it is fitting to the space theme more than wooden pieces would do.
jochen tells that he doesnt like to play terrans, although theyre very strong but sebastian do like them.
jochen thinks firaks are one of the strongest and are very fun to play although they have only one available strat (often up- and downgrading the labs).
his favourite faction are xenos (not because they are strong but they are fun). he does enjoy starting with a third mine and with his standart qic-academy in first turn he has early techtiles via qic-actions and he often makes many federations with xenos.
jochen doesnt like how lantids are way too weak. he does think one major problem of them is that they start with only four powertokens and those are all in bowl 1 at the start of the game. its the only thing hes doesnt like about gp. he talked with helge ostertag about that and helge told that lantids were to strong in playtests and were therefore nerfed very hard (and probably too hard).
jochen thinks that its a pity that there will be no expansion for gp. he would really like to see some new tiles or an additional 2 factions per planet-type. lines doesnt know any news about that because he has no contact anymore to feuerland-spiele.
sebastian would like to see something about backroundstory of the factions.
  • [+] Dice rolls
Jack Spirio
Austria
flag msg tools
badge
Avatar
mbmbmbmbmb
Great summary thanks
But it’s power not might
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Jonas Egel
Germany
Solingen
Deutschland
flag msg tools
Jack Spirio wrote:
Great summary thanks
But it’s power not might


Oh im dumb. I even knew that. blush
Ill correct it in the text, thanks.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Sebastian Schmieder
msg tools
badge
Avatar
mbmbmbmbmb
Wow, that's awesome! Thanks Jonas for doing the job I was too lazy to do :-)
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Jonas Egel
Germany
Solingen
Deutschland
flag msg tools
Basti725 wrote:
Wow, that's awesome! Thanks Jonas for doing the job I was too lazy to do :-)


Youre welcome. This way I had a chance to thank for the nice podcast
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
James Ataei
United States
North Carolina
flag msg tools
designer
badge
Avatar
mbmbmbmb
Thank you very much for translating this for everyone!

1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
JGS
Germany
Heimstetten
Bayern
flag msg tools
Trank you Jonas very much for this great recap. Thank you also for praising this podcast. It was fun to talk with Sebastian and Linus.

I would like to say thank you to all other GP fanboys for your great postings, especially in the strategy forum. I would like to highlight Lenrok‘s postings in the high score Thread on 11th February 2018 and 17th February 2018 with pictures. They were awesome.

It‘s a pity that playing with you in PBF is to complicated for me. But I hope there will be a posibility in the future to play with guys like limitless, Kester, Jack Spirio, frotes and the other GP fanboys. The Taklons opening in PBF 9 from Kester was great, just as the Geodens victory-way from frotes in the same game.
Greatings from Munich, Germany.
Jochen
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Jack Spirio
Austria
flag msg tools
badge
Avatar
mbmbmbmbmb
Gaertner3031 wrote:

It‘s a pity that playing with you in PBF is to complicated for me.

It's easier then it looks, give it a try
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Prev «  1 , 2  | 
Front Page | Welcome | Contact | Privacy Policy | Terms of Service | Advertise | Support BGG | Feeds RSS
Geekdo, BoardGameGeek, the Geekdo logo, and the BoardGameGeek logo are trademarks of BoardGameGeek, LLC.