Randomness in early game

Have you ever watched professional poker? That game is all about taking advantage of the randomness.

The randomness in PA will lead to richer strategies. Instead of 1 ultimate build strategy you will need something that can adapt to the situation. You won't know where your opponent has spawned so you need to defend all directions on your base instead of just one.

I think the biggest problem with your scouting example is that it demonstrates poor strategy. Your opponent was only able to get the drop on your engineers because you left them defenseless, gambling that they wouldn't be found. You need to make a calculated risk and either protect the engineers or maybe send each off in two different directions instead of putting all your eggs in one basket.

unfortunately, poker has what matiz insists on. 'Mathematical fairness'. This game however does not.

Personally, I'm extremely glad that PA has abandoned the 'learn best build or die' mantra adopted by other RTS games. That sort of mentality is a massive barrier to new players and as others have pointed out in this thread already, leads to repetitive, boring games.

I don't particularly want Uber to remove randomness from the game for this exact reason, and I'd much rather we embraced the different mentality it forces - learn the game mechanics, and how best to use them to your advantage in a given situation. Not rote-learn a plethora of pre-designed builds.

For competitive games my preferred solution would simply be to use 'best-ofs' to average out the randomness.

This game definitely needs some randomness, but some randomness like metal spawns and actual spawn locations need some work. Unfortunately i dont think the procedural generation that we have will ever fix it, and the only way to keep procedural generation in tournaments (so that every round is not a bo5 or more) is by limiting some of its randomness. Two halves of a planet being symmetrical does not remove any randomness from the game, but people are against it even though it would bring about fairness with metal distribution.

A lot of this is probably tied to metal density, currently there's far too much on any given planet and excess is distributed at -- like everything else -- at random. Asymmetry means that one side might have to go a little further for an expansion mex..maybe, but lets not forget the planet is round and there are numerous expansion mex in just about every conceivable direction, odds are at least one is closer to the other player. Symmetry is not needed for fairness, though an even number of metal per spawn location -x if terrain feature is present would go a long way (not sure if possible or if this is the intent, would be nice though).

The problem with this is that we cannot have high density metal clumps. And there arnt metal expansions in every conceivable direction because not only is metal distrubution completely random, but terrain features that block metal points are too. Kinda unfair if one player can only spawn in mountain biomes and the other has plains.

I'm not against having a more 'balanced' algorithm for generating tournament planets, but I'm sure that would involve either a lot of work on Uber's part or a particularly skilled modder's dedication to re-working it - that's my only reservation.

Something will definitely need doing for tournaments, and the likelihood is that it will be needed for ladder too. We cannot know the difficulty until we see Ubers implementation of it/the game, but i dont see why difficulty should be a deciding factor in this.

You do get several spawn location options, usually what 3? How often do all suck exactly? How likely is it that your opponent(s) have the same or similar choice to make at the start? It is quite likely that everyone as at least one bad spawn option and at least one good one. This is already reasonably fair, its down to which starting spot you choose and why so its really not as unfair as you think. How many games do you feel you lost/won purely because of luck...not many I wager.

There is a realistic change to get really bad spawns I'd say. I have no numbers, but it is too high. Sure, if Uber improves the map generation a lot it could turn out really cool.

How is poker fair? That game is based on luck and randomness, even with all the card counting etc.

because we do not take into account how good or bad a player is when deciding if a game is fair. Poker is mathematically fair in tournaments because a randomly shuffled complete deck is used with every new hand as far as i know. Its fair because there is an equal chance of drawing any card. Its up to the skill of the players to then win games.

It would be fine if this game was mathematically fair, but it is not.

It should be 0, and not 'not many'. As people get better at games, balance becomes an issue. Uber have avoided the whole catastrophy of race imbalance, but they have in turn introduced map imbalance instead. Sure, over a large spread of games it will even out, but every game matters.

It is fair in the sense that all players have the same probability to get a card, yes.. but getting a random card from the deck is not based on skill whatsoever. The skill is in the probability management..luck is still a big part of it. It evens out over many rounds and then the skill starts to matter more.

Just as well you might say PA is fair, all players have equal chances in spawn/scout luck in the early game. Thats why you play BO7 etc.

strange, i thought the skill in poker was the ability to read your opponents cards and make your opponent think that he knows the value of your hand. The only time in poker that the hands really matter is when two players go all in, but even then, the fact that they go all in is based on their skills of thinking that their cards are better than the opponents.

And as unfortunate as it may be if you are wanting to play BO7's against people in ranked matchmaking, i do not. The posibilty of 7 games in a row just to make a ranked match fair is just plain wrong. The same stands with tournaments.

Yeah if every tournament introduced bo9 or at least bo7 randomness wouldn't be a problem, because it would even out mathematically just like in poker. Any kind of ladder match would also have to be bo9 or bo7 to secure fairness. Randomness can be overcome by introducing more games to even out chances and probabilities of players.
Unfortunately:

Even starcraft which is faster game than PA rarely introduces anything more than bo3. This is not the solution, too much time.

Ladder is about the big picture, averages and consistency of performance anyway, you play hundreds of ladder matches, so that does even it out. You don't need bo3 or bo5 or bo7 in ladder. BO3 in tournaments does hit the sweet spot in many ways I suppose, and is not that tiresome to watch and play.

I guess this discussion comes down to what one is looking for.. if you are looking for a competitive sport, absolute fairness is a big deal. If you are just playing a light war simulation game for fun, you don't mind small random elements.

I agree.. though I feel metal point distribution got a bit worse and uneven with the latest patch, it was better before. Might be just my imagination though.

My two cents :

I don't think that selecting a spawn is a good idea, at least as it is now.

Currently, out of the 3 proposals, given you don't know where your enemy is exactly, there is only one good spot :
The one that allow you to get the biggest eco the fastest.
Choosing another spot is an handicap.

That bring two problems.
Some people seems to think that that choice make the game rely less on BO. I think the opposite : It make having a perfect BO more important, because :
- You may have a slight disavantage on start, so any bump in your BO put you at a greater disavantage.
- It's the only thing that you can do in order to win (ie. good luck winning with a fast raiding like in supcom).

Second problem is :
You have to learn how to recognize good spawns from bad spawns.
That's something have to know and can't really learn from experience :
Did you lose because of the spawn, your BO or a bad decision? You don't know if you don't know how to choose the right spawn, do a perfect BO and play perfectly.

That's a case of depth vs complexity : It make the game more complex to learn, without bringing any depth (as the spawn choice is an illusion).

And sadly, in my opinion, PA is full of complexity but has no depth.

I don't think "the egg" will change that :

- We don't even know what it is.
- Even if it works great, there will still be a best option (I don't think it's the kind of thing you can render equal, unless proven otherwise), so it just add another layer of complexity, as well as a new thing to learn before thinking of playing the game.

All in all, I believe that procedural generation is/was the biggest mistake of PA :
- Maps are not interesting. Sure we have an infinity of them, but it's basically the same (unbalanced) map with some variations. Also, no interesting chokepoint or any terrain situation advantage.
- Maps are unbalanced. Making them balanced will reduce randomness, and so will not justify procedural generation.

Uber has a nice brush thingy for creating planets, I would have hope that it was the editor instead of what we have now.

Another possibility would be to create random planets based of interesting patches of planets :
You create manually quarters of planets, and the generator mix/scale them up, probably mirrored. And tadaam, best of two worlds.

At some point, Uber should stop trying to be different, and try to be efficient. Others games work great for a reason, let's start from there.

great post I agree

This is also pretty important question which I've slightly touched before on earlier pages when I was descrbing 3 steps that we need to do in order to make game better competitively. It's just way harder to actually practice because of this.