Am I the only one who never won a game?

[Num7]

I've been playing FreeOrion for quite a while without actually experiencing victory once. Not even once. I'm used to how things work game-play wise, but I can't get to stop the AI from invading my empire with gigantic and unstoppable fleets when things get more serious.

I always get my ass kicked by the "Beginner Mode" AI.

I'm just out of a 1 on 1 game against the easiest AI, in a 10 systems galaxy. I had 5 planets while the AI had only 1. I thought I was doing okay, until I was attacked by a 30 vessels fleet around turn 70. They captured all my planets and destroyed my small fleet. There got to be something I'm not doing right.

Can someone give me a few tips and tricks regarding the early stage of the game?

Thanks for helping me!

[MatGB]

Early early stages, I build scouts to make sure I've got the whole map discovered and hopefully to watch the whole map when that's done. I didn't until recently tend to build any warships until I'd got a decent hull and a decent weapon researched, but more recently I've been building small mark 1 ships very early as a swarm of them can take sentries and forests.

Early priority is the first learning tech that gives you a research bonus, then Subterreanean Hab as that gives a bonus to all outputs (the Archives on your homesworld is a bonus for production based on population, more pop= more output even if focused on research, which is what you want to be doing until you can build something worth having). After that nascent AI as that's an obvious Good Thing in the early stage.

I rarely build colony ships, merely using them to claim key strategic resources and/or to extend supply lines, it's far easier to invade someone else's colony which is what I do.

Concentrate on one hull line, either asteroids or organic/xenocoordination. If playing Etty or Eaxaw, or you get native Mu Ursh or Hhhoh early, then you can get away with mass driver to 4 and stay there for ages, if not you need laser 3 or 4 early.

Once I've got Organic Hull, which is my next tech after AI, the incubator is top build priority on homeworld (or best conquered native), and I research a mix or laser, zortrium plate and active radar, normally L1, radar, L2, zortrium, L3. Depending on research output that tends to complete at roughly the time the incubator is ready. Then switch homworld to industry, spam ships and go invade. Never, Ever put shields on a ship that has a build cost without them of 200 or less, simply not worth it, on an Organic Hull, 2 armour, one weapon, extra fuel, job done.

The best way to ensure the AI hasn't got a massive fleet in orbit around your homeworld is to have a massive fleet in orbit around their homeworld first

[jenschou]

Another genre of tips:
Choose a Setup that the AI is bad at, most notably that would be sparse galaxies, i.e. low amount of planets and starlanes.
Play around with the Galaxy Setup a bit. Choose a race that is strong for that setup.
If you choose a very small Galaxy, like the one you described in your post, then the best race for you is one with good pilots, like the Eaxaw.
In a larger Galaxy, Gysache are the easiest race (but then you should ensure that the 'Native Frequency' is not low, so you can easily get good pilots).
You need to adapt your playstyle according to what setup you have. If you are on a small map, you will meet the AI early, so you need to have ships early. If you are on a large map, it will take much longer until you meet the enemy, so you should get your empire up and running first.

[MatGB]

Good point, yes. Easy mode: Chato on Young.

Chato have excellent research output, but don't suffer for industry, and on a Young galaxy you get lots of bright stars, their broad tolerance and phototrophic metabolism means they can colonise a lot of places without any research, and get higher population on wherever they colonise.

Cray are also nice and straightforward on any galaxy age, but suffer industrially. George take a bit more skill as they lose research points. Don't, ever, play Egassem until you're really confident, they're truly terrible with very few redeeming features. Laenfa are interesting but with some weird quirks—on the other hand, you're basically invulnerable when playing them as the AI can't see your colonies to invade for ages due to their stealth bonus. Etty are weird, Trith annoying (the xenophobia stuff makes them very tricky and their advantages aren't as obvious as the Eaxaw). Gysache are potentially incredibly powerful but the Bad Pilots thing is a drawback you need to work around which I'd avoid while learning. But if you have High Natives as Jens says they can be brilliant.

But yeah, Chato on Young Galaxy, with a bit of breathing space to explore.

[Num7]

Thank you for your good advice.

I was able to win my first game tonight with the Chato on a young galaxy of 48 systems. I was a bit worried at the beginning because I was able to see an enemy fleet building up a few systems from my empire. But I was able to upgrade the defense attributes of my border planet facing the enemy territory enough to repel their fleet, even those few times when they were able to beat my fleet, they wouldn't stay in the system. I assume they left because of my defenses, right?

Almost all of my planets were focused on research, except a few in a system with a gas giant which had a generator. That system generated the production points for my whole empire.

At some point, I dropped the research and started producing tons of powerful ships, building a massive fleet. I got to conquer the entire (48 systems) galaxy.

I still have trouble deciding if X or Y planet should have a research or production focus, but tonight's game gave me an idea how to balance those 2 important concepts.

Thank you for your advice up to now. I'm all ears to find out about other strategies and other tricks to master and experience FreeOrion at its best.

[Nick]

I've also had absolutely no success winning any games with the current build either (0.4.3 sv 7114).

It seems like no matter what I do, the AI steamrolls me with huge fleets of ships extremely quickly - even with only 1 opponent and AI set to beginner mode.

Basically.... beginner mode is too hard. Drive a spike into its brain and deprive it of oxygen for about 5 minutes, then it might be a little more "Beginner".

[MatGB]

The problem we have here is the game has a steep learnign curve, but once you've learnt how to win (it took me a few months and I used to die horribly regularly) the game actually becomes quite easy.

It's actually quite hard for me to give advice on the whole "how to survive the early game if you don't know what you're doing" thing because I can't remember how hard it was and I'm, now, constantly looking for ways to make the game vaguely challenging, I haven't not won for at least 6 months (abandoned several games out of boredom, and one or two as the layout was clearly stupid, but never been defeated).

And that's on Maniacal setting.

But yeah, general rule of thumb: all planets should be set to research early on unless they have a gas giant or, at a pinch, an asteroird belt or have a special that improves industrial output, even then I'd consider research if the species doesn't have a bad research malus.

However, once I've got the ability to build effective warships (ie laser3, organic hull, zortrium armour are all researched) I tend to switch a couple to production, especially my homeworld in order to spam out a fleet—the best design I have set on repeat, increasing the number to be built once my outpuat has risen enough.

Try not to worry too much about having fleets everywhere to defend, use scouts to watch for incoming, concentrate your forces and go on the attack whenever possible: three ships plus a couple troop ships is normally enough to take an enemy in the early game.

Marc, to answer your direct question, no, probably not, once an enemy fleet is in orbit, planetary defences will have been destroyed, staying there with a couple ships is relatively painless. why they'd have left I don't know, possibly they couldn't bring troop ships up, there was another threat you couldn't see, or the AI just forgot what it was doing and wondered off (it does that sometimes, a lot less often than it used to but…).

But it's good to know the advice has helped. I think we really need to work up a strategy guie to replace the horrifically outdated one on the Wiki for 0.4.0, it just requires having the willpower to stop playing for long enough...

[Herode]

But yeah, general rule of thumb: all planets should be set to research early on unless they have a gas giant or, at a pinch, an asteroird belt or have a special that improves industrial output, even then I'd consider research if the species doesn't have a bad research malus.

I don't agree at all with this rule.
I've been playing FO for a full year now, let's say I've played a dozen of games. Most of them were interrupted by early defeat, let's say between turn 50 and turn 100.
All those games I've played with policies giving high priority on research (through planet's focus and/or choosen race).
I'm now playing #7136 with George (bad research), thus focusing on Industry and mass production.
With AI set at "typical", it's right now the best position I ever had at turn 100 : 41 vessels, 12 planets, industry 166, research 68.3, 167 pop.

I guess several factors have to be kept in mind when you play :

- focus on your race's strength. Good research ? Go scientist. Bad research ? Go mass producer. For my current "George" game, I didn't even bother to search for Algorithmic Elegance before I captured my first ennemy planet, uninhabited by Etties. I had 5 Research Points by turn until that moment (t55). But enough ships to capture the ennemy capitol.
Then, and only then, did I started to colonize other planets with Etties colony ships and turning their focus to Research.

- never never never (never...) let your capitol without defenses : AI empires love sending ships in our home system, this will cost you a lot of ships/time and eventually, the game.

- be a bad guy.
Foreign scouts ? Kill them all ! Foreign ships ? Kill them all !
AI despises shy ennemies but is very cautious when it comes to close a rabbid dog.
Don't forget : YOU are the bad guy.
How bad ? REAL bad !

- your chances to win a game depend on your starting position, that is to say : on the seed and the galaxy configuration (age, size, shape).
Some starting positions will be very difficult for humans but easy for Georges, for exemple. If you don't have colonisable planets within 3-4 systems from your homeworld, then the game will be real hard, no matter if the AI is only set to Beginner or Turtle. Thus : adjust your race from the starting position if you want to tune the difficulty of the game.

- wisely design your ships, use decoys, don't forget shields at t100+. Read the tech's description, it is consistent with what it does and what fields it opens.
When you gain a tech, really use it otherwise you've just spoiled points. Spoiled points means spoiled time and eventually lost game.

IMHO, there are definitely several ways to run a game and win and.
I think that's good news : it would be boring to have a single deterministic path to victory.

[Goemoe]

Basically.... beginner mode is too hard. Drive a spike into its brain and deprive it of oxygen for about 5 minutes, then it might be a little more "Beginner".

This. It is simply annoying to get beaten time and time again on the easiest of settings. This should be adjusted, if you want to attract more people.

The problem we have here is the game has a steep learnign curve, but once you've learnt how to win (it took me a few months and I used to die horribly regularly) the game actually becomes quite easy.

The point is: there are a lot of people who don't learn for themselves how to win on the easiest of settings. Beginner should be just that: a seeting for beginners. Once they have tried this or that, they should beat the game every time getting bored of the beginner setting. But FreeOrion has no such 'Beginner' difficulty only a setting named beginner, which it is not. You write (it took me a few months and I used to die horribly regularly). For people interested in this game, this simply should not be the case on beginner.

[Vezzra]

It is simply annoying to get beaten time and time again on the easiest of settings. This should be adjusted, if you want to attract more people.

Which version do you use? The latest stable release (0.4.3) or one of the recent test builds? If you're using the 0.4.3 stable release, this one is seriously outdated, and AFAIK our AI programmer (Dilvish) has nerfed the AI on easier settings already in repsonse to complaints like those in this thread. But these changes are of course only in test builds released after 0.4.3.

So, you can either grab the most recent test build, or you wait a bit, as we are currently preparing to get out the next stable release, 0.4.4. We plan to provide the first release candidate build coming Monday.

[Foocaux]

Wow, it just goes to show the disparity of skill level the devs need to cater for!

For my (first and only) 2 games I had the AI set to maniacal, and I was already wondering if there was a way to make the maniacal AI more manic somehow. I had a couple of hairy moments in both games around turn 200, but luckily no competing empire did build/research any mini asteroid swarm ship, otherwise I would really have had a real fight on my hands...

I s'pose I've played way too many 4x games in my life...

Back to nerfing the beginners AI, if it's still too difficult in the current test version, in what way should it be nerfed?
It's very much possible that the developers - given their familiarity with the game - find the beginner AI so easy to beat already that they can't think of a single thing to make it less of a challenge.

Maybe letting them have less production & research points?

It's just a thought, but only giving 90% of the current PPs & RPs to the beginner AIs might already make enough of a difference.

[MatGB]

Wow, it just goes to show the disparity of skill level the devs need to cater for!

Pretty much, I regularly look at some of the complaints, or even in discussion with experienced players, and my brain blanks, to me some things (like, for example, high number=good, all other numbers based on population ergo high population early= very good) are just basic and obvious, to others they're a strategic choice and they choose to do somethign else first that I suspect would happen quicker if you got the high numbers first.

Anyway, rather than start a completely new thread, as it's relevent to here, over the next couple weeks, in preparation for the full release of 0.4.4, I want to update the quickstart guide:
http://freeorion.org/index.php/V0.4_Quick_Play_Guide

It was written for version 0.4.0 and is horrifically outdated, but rather long and contains some good information. If anyone reading this has some suggested edits, or notices a specific section that's definitely either wrong or worth keeping/updating, post here?

In addition, after 0.4.4 is released, one of my plans is to code up and introduce a series of "tutorial" sitreps to give players helpful advice as the game goes on, these will, hopefully, link to Pedia entries with some quick,b ut helpful, context related text. That's not going to start happening immediately but if people have ideas or stuff that, specifically, confuses them that would be helpful.

(for example, turn one will definitely include advice to send the scouts out to explore, avoid monsters and plant your colony on the first habitable world you find, all of these are "obvious" to me but some just don't do it)

[Geoff the Medio]

It's just a thought, but only giving 90% of the current PPs & RPs to the beginner AIs might already make enough of a difference.

The game does not treat AI and human players any differently. Everyone's playing by the same rules. AIs are free to be dumb, but the game engine is not going to apply player-specific penalties of that sort.

...introduce a series of "tutorial" sitreps to give players helpful advice as the game goes on, these will, hopefully, link to Pedia entries with some quick,b ut helpful, context related text.

The alternative / next step beyond that is to implement in-game "quests" that prompt players to do stuff they should / could be doing to advance their empire, and give in-game rewards (techs, ships, information, bonus sources of PP/RP, whatever else becomes possible) for completing them. Such "quests" would hopefully serve actual gameplay purpose though, and not be just informational / for beginners.

[MatGB]

Works for me, but I suspect it'd need some backend coding and far more work, a quick "welcome, try and do this" that could be coded fairly easily (now that I've understood more how turn based triggers work in scripts).

We definitely need to do more to welcome players in, but a quest might require ensuring there are specific things in there and that might be harder.

[Vezzra]

Works for me, but I suspect it'd need some backend coding and far more work

Well, when I started to implement scripted universe generation I already had the vague plan in my mind to make the entire interface available not only for universe generation, but also for Python scripts which can be executed at the begin and/or end of each turn. That would give tremendous flexibility for creating things like scenarios, and would also provide the means to implement such things like "quests".

As the interface is already in the server, it shouldn't be too hard to add something that use it to execute Python scripts every turn.