# Starry Skylines (Part 3)

devlog

This is part 3 of my devlog about the game Starry Skylines!

If you haven't read the other entries, click this: [Devlog] Starry Skylines (Part 1)

## Intermezzo: bigger is not always better

By far the biggest issue though (after some more quick playtests and careful consideration), is the sheer number of possible buildings and effects in the game. I need to find a way to make it slightly less random and more likely that you can use combos/follow a strategy/have some idea about which buildings to expect.

For example, I've had one game where 6 or 7 "wheat fields" appeared. However, they only work in combination with Farms/Bakeries. On their own they do nothing. Guess what? Absolutely NO farm or bakery appeared, even though they have quite a high probability of doing so.

This means that it becomes hard to plan ahead and actually strategize and build something nice. Many of the things you do ... will just not have any effect (if we're unlucky with the random generation).

The first solution is to alter the probabilities. Now many buildings have roughly equal chance of being picked; I should change that so buildings that have a greater effect (or belong together) have a greater probability of being picked.

The second solution is to prune some of the tree. Remove the buildings that are least fun, only keep the best ones.

But all those solutions are imperfect, as they can only soften the problem, so I'm considering the following strategy: include "combo"-information on the buildings and events.

Whenever I show a certain component, I remember that information. The next time, I increase the probability of picking something that is somehow related to this previous component. This shouldn't be a certainty ("oh, we have a police station this round, surely some criminals will pop up next round"), but it should increase the probability that options you pick (and events that appear) actually influence each other a lot.

I don't know how feasible or helpful this is. I've never done it before. But I don't see any other solution: all buildings/events/themes/planets I've created so far, are really nice and interesting. I'm happy that I could create such a big list of diverse mechanics and situations, but with every playtest I see this huge size as more of an obstacle than a feature.

(I already increase the probability of buildings of the specific planet you're playing, which helps somewhat. If this is your first time playing Intervenus, its unique buildings (such as Police Stations) show up way more often, allowing you to get comfortable with them.)

## Intermezzo II: for sure, bigger isn't better

After thinking about it some more, you could of course ask another question: should this even be one game? The mechanics are so diverse and widespread, shouldn't this just be multiple games?

I think the answer is: no. The core of the game stays exactly the same between planets, and I don't see any significant changes I could make to reasonably split this into 2, 3 or maybe even 4 games.

Instead, it's probably more helpful to think in the other direction: if this must stay the same game, how do we make each planet feel unique and coherent within the whole game? And then I'm starting to think I need more options/settings for the game.

Instead of picking a planet and getting all previous planets as well, you can cherry pick which "expansions sets" to use. Maybe I could even create something similar to "decks" in those trading card games. One of those decks could be "Nature", which includes the planets Marsh and Pluto, as they are the only ones strictly focused on nature. Another could be "Leadership", which contains Uronus and Intervenus, both focusing on government and planet leadership stuff.

Hmm ... hard to make an informed decision on this one. I'll just keep polishing and 100% finishing the current game while thinking about a solution to this whole thing.

## What I ended up doing ...

I ended up implementing a mix of all these ideas.

I would've loved to implement an algorithm that picks random buildings that are somehow related, but it just wouldn't work.

Why not? Let's say I create a set of 20 unique buildings. And, by chance, only one or two of those use the resource grid (and the rest doesn't).

Then I would have to include the resource line component, and it would show up quite often, even though it's only rarely useful. Which is annoying. It skews probabilities in the wrong direction, and I can't do much about that.

So, the safest bet is to create random sets of complete "planets", and limit those sets to at most 3 or 4 planets. This way, everything will surely be related, and nothing shows up way too often or way too little. Yes, again, I would've loved more variety and randomness, but this is the only way I saw this working out.

Lesson learned for future games: think about the problem of quantity beforehand. There is such a thing as too many options or too much stuff in a game, which causes a game to lose focus and interconnectivity.

## Fourth Playtest

So, after some more rounds of playtests, I've made these observations.

First of all: this was the first playtest with a starting setup generated by the computer (I had just coded this part of the website) and proper planet sets. Fortunately, all of that works great! No need to tweak that very much.

Issue #1: currently, when you play with 2-3 players I have the rule that you can't choose an option that another player already chose. That works really well, which is why it's a shame that I throw away this rule on 4+ players.

Additionally, it's a bit confusing that there are three options per round, but there can be four players on a planet.

So, I changed the rule to be more general: "each planet starts with 3 players, each of which must pick a different option" (Way simpler, more logical.)

Issue #2: there were some buildings that showed up, which were ultimately useless. For example, in one game we had loads of Paddy Fields ... but they can only be built on water, and we had no water, because our set of planets did not include the "water planet" (which is Marsh).

How to fix? These kinds of buildings are, fortunately, the exception to the rule. I can add a parameter (in my gigantic list of buildings) that says "requiredPlanet = blabla". Subsequently, the building is only included if the required planet is available.

Additionally, some buildings might just move to the first planet (the "core set"). They are so useful and interesting, that it's a shame to hide them on another planet.

For example: one cool feature of this game is that you can travel between planets. But the only building that allows you to do so, appears on Uronus. Which is a shame, because that planet will not be included in the set a large chunk of the time.

In conclusion: I need to take a hard look at my first planet and make sure it includes all the best bits.

Issue #3: in this game, you can place stuff anywhere you want.

In the first version, the rules were way more restrictive. You could only place a building/street if it was adjacent to another thing you owned. Although that worked fine, it wasn't much fun, and it was easy to lose the game halfway because you just didn't have any connections left.

I completely removed that rule, which was an improvement, but still isn't perfect. There's often no reason to plan ahead or create beautiful street networks, because ... well ... anyone can block your plans at any time, and you can do the same to others.

I need some rule that restricts placement, but only slightly ...

## Better Player Powers

At this point, I remembered that my idea for "unique player powers" was also broken.

(I had 8 different powers, which were unbalanced, hard to understand, and sometimes worthless unless you were playing one of the later planets. Yeah, these relics were left over from the very first draft of the rules for this game :p)

So, let's do the age-old trick of combining problems to get a single solution!

I want every player power to have both an advantage and a disadvantage.

The advantage can be something like: "you start with one more path than the others" (which also nicely breaks ties for start player).

The disadvantage is something that restricts placement: "but you may not place buildings next to paths owned by someone else!"

What am I going to do now? I'm going to think about these player powers, fix the rules, add some more parameters and balancing to the huge list of buildings, and perform another playtest session. I think the game is nearing completion, but I must also say I'm growing tired of this game. (As I'll explain below, I created a game in a genre that I don't even like myself, and it's become way too big. So that was a recipe for motivational disaster :p)

Why do we even need player powers? Without player powers, the game is completely symmetric, which means there's no fair way to decide a start player (or general turn order) and everyone can mostly copy each other's turns. Additionally, I've learned that it's always good to give players a sense of ownership/customization: by choosing which "role" they want to play, they become more engaged with and interested in the game.

## Final version?

So, all elements are done and working, the game is as finetuned/balanced as I can currently make it, time for the fifth playtesting session.

It went well! (As the heading of this section already spoiled, kinda.)

It's hard to playtest this game well, on my own, as it's just so large. There are 8 planets, over 100 buildings and effects in the final game, and this game is best played as a campaign (where you play through all these planets with the same group).

As such, I'm most uncertain about the later planets and the game as a whole, but ... the first few planets are certainly fun, easy to explain, easy to play, and quite addictive as well!

It's like I said: this game is nothing more than presenting three interesting choices each round.

I must say, though, that I have a slight feeling of ... disappointment. But that's not the game's fault, I've realized, as it works great and people have a lot of fun. (It's actually the favorite game out of all my games for some people I've played with.)

I think it's because I've never been a fan of "point salad" games. And this game is almost the definition: you score points for all sorts of things in this game, and you only know who won by performing a few minutes of calculation at the end. It's the first "point salad" game I've ever made, and I highly doubt I'll make one again.

## Conclusion

This is probably my "biggest" game yet, and also the game that is most unlike anything I've ever made before. The fact that it's even playable and actually quite good, means it's a success for me!

I've planned more balancing and playtesting sessions in the near future.

(I actually do this for most games: plan a date in my agenda, a few weeks or months later, to play the game again. Usually reveals quite some flaws and polishing issues that I was too blind to see when I was full-on developing the game. And, more playtests are always a good thing!)

Additionally, I've written quite an extensive to-do list for possible future improvements/changes. I'm just never really satisfied with my work (which is both a blessing and a curse), so although the current game is fun and polished, it might get a big overhaul at some date in the future.

Hopefully you learned something from this devlog, and hopefully you have fun playing this game!

Until the next devlog,

Pandaqi

• «