# How do you decide on the size of a tech tree?

Is there any established metric with which a game designer might decide the best size for a tech tree?

In this case the game is a single player 4X game (thank you, Peter, for that term). I wanted to incorporate an appropriate length and breadth of tech tree as a basic design goal.

For instance, is there data on player reactions to tree size vs game scope or game length? Or even a reasonably established rule of thumb?

• I find that game design questions generally attract more concrete, focused answers when they define the context of the game you're developing, how the feature works so far, and the player experience target you're trying to tune or modify it to achieve. There are precious few game design rules that hold unmodified for every version of a feature, so the more precisely you can define the creative needs & aims of your version, the better we can target advice to it. See the game-design tag guidance here. – DMGregory Dec 6 '17 at 21:10
• "tech tree" remains an obscure term to me. Am I right thinking of the logic that guide "you can now research on X, since you have developed Y" in e.g. Civilisations ? – sylvainulg Dec 7 '17 at 9:45
• @sylvainulg Yes. – anna328p Dec 7 '17 at 19:29

What you must worry about is the number of choices in the tech tree.

In terms of a 4X game, lets say the depth of the tech tree is what affects how long a single play-through takes, while the width of the tech tree affects how many choices a player has each time they pick a new technology. The depth is determined by your game design, but the width should be limited, and relatively constant over time, although it's common to start narrow then get slightly wider as the game progresses.

When letting the player pick between different options, you should generally try to keep the number of options between 3 and 8* If you want a wider variety of choices than that, a good way around it is to implement a hierarchy of choices - for example in MoO you could chose from 8 fields of research and in each field you could chose from up to 3 techs. In Crusader Kings there were 3 types of techs with 8 sub-choices each.

Figuring out how deep you want the tech tree to be depends on too many factors to be solved with a formula. But in most cases it's at least 3 levels deep, so I say implement 3 and start playtesting. Keep in mind that more than 10 levels are quite rare.

* Why 3 and 8? 3 is the minimum required for a multiple choice scenario, which players experience slightly different from a binary choice. 8 because it is believed by some that 7 plus or minus 2 is the upper bound of objects that can be retained in working memory - more choices than that are very likely to overwhelm the player making the choice and cause them to just pick one at random. Truthfully, I just wrote 8 because I know of some games that set the upper bound at 8, but in practice I like to aim at 5 myself.

• I'd say this is not bad, but the numbers aren't all that accurate. Two examples that come to mind are the Sims games and Elder Scrolls V: Skyrim. Skyrim has 18 skills, with anywhere from 8 to 15 perks per tree. The Sims does one or two sub-branches per career tree, and a single tree for skills, with numerous different trees. From personal experience, as long as the progression through the tree is made clear, and read the game's audience. Hardcore RPG fans (like Skyrim) are more likely to read through each skill for min-maxing than those in the Sims, where the game is much more casual. – Anoplexian Dec 7 '17 at 16:22
• @Anoplexian Could you elaborate? Your example, Skyrim, has 3 skill groups with 6 skills each. The skills allow you to select from 1-5 perks each skill level, and the selection is 4-5 levels deep. – Peter Dec 7 '17 at 16:28
• I wouldn't consider them to be skills groups as such, simply because none of them are exclusive from the other. For instance, I can be proficient in every single on, but when we go down a level for the actual skills, it provides a locked tree in which the perks and levels don't interact. Since there's no interaction, and since each skill level has many more than one path, I'd say the trees are the skills themselves. It's not like 3-8 is a bad number, but I'm of the impression that a more involved game such as Skyrim has more involved players who can more enjoy more involved trees. – Anoplexian Dec 7 '17 at 16:35

In order to answer this question, we first need to consider why we have tech trees in the first place.

• They limit the complexity in the beginning of the game by hiding the more advanced game mechanics. This makes the game more accessible to beginners.
• They serve as a reward mechanic. Unlocking a new tech tree node gives the player a feeling of progression and accomplishment (as long as they are relevant).

The size of the tech tree is dictated by these two considerations.

First you need game content and features which are worth unlocking form a design perspective. What to add and what to leave out is an important decision which is influenced by both design considerations ("Would adding 20 more unit types make this game more interesting or just more confusing?") and business considerations ("Yes, it would make the game better, but do we have the budget for it?"). And in what order and frequency to unlock content and mechanics to make sure every phase of the game is interesting to play is another important game design decision. Both of these design aspects are far more important than the UX of your research screen. And there are no simple rules of thumb for them either, because they depend too much on design details of the overall game concept.

(Not that the UX of your research screen doesn't matter. But its job is to sell the complexity to the player. You shouldn't optimize the complexity of your game solely for being easy to represent on the research screen)

And then you need to make sure that every tech tree node is relevant in order for it to feel like a reasonable step of progression. "Now I can build space ships" is a relevant progression step. "Now my woodcutters are 0.15% more productive while it rains" is likely not. You also need to make sure that each tech node is useful just by itself. Unlocking railroads, locomotives and train wagons separately when none of them serves any purpose without the other two is just busywork, so they should be merged into one tech node. This limits the number of nodes into which you can divide your mechanics and content.

That means the size of the tech tree is determined by the depth and breadth of your game mechanics and content, not the other way around.

• In addition to this, it's always a good rule of thumb to have more choices in your tree then the player can make. This can put more weight to a choice, and might increase the replayabilty – sjoerd216 Dec 6 '17 at 14:18
• Well the tech tree in Rogue's Legacy is a good counterpoint: you were eventually able to unlock every branch, but it was still engaging through multiple playthroughs. What put weight on choices is when you unlock branches; since you were using up a limited resource (gold in that case) you had to decide what order to unlock branches. – jhocking Dec 6 '17 at 20:27

I think you should start from definition of the context: which kind of games? genre? single or multiplayer?

In my experience as player and hobbyist, less options with meaningful effect into gameplay are desirable instead of a lot of options with few impact on gameplay. And keep in mind that presentation to player is a critical point. A huge (100+) tech tree is hard to fully master for a player and could be difficult even to display. Also is not clear how much important is the tech tree, I mean if right choice is critical in your game or is something not even mandatory to win.

Some examples:

1) less options in a tech tree, like in Starcraft 2. Because the realtime and mostly multiplayer nature of the game, game designer decided to stay on a limited choices. Games duration here the critical point: a SC2 game in average could last 11 minutes. In that time player has to control everything, understand and use tech tree and win.

2) more options like in Hearts of Iron 4. Because game is mostly single player could be paused and games could require a lot of time (20 hours?) many players stay a lot of time inside the game, game designers develop an huge tech tree ( more than one actually). Mastering a tech tree in order to win is critical and many guides are written only on this

• You can also restrict player from some branches(nodes) if they pick another path of development which would ease the pressure of player choice. He would know that this branch or node is better for him to enhance his current position and starting a new branch would be a waste. Civilization Beyond Earth does that. – Candid Moon _Max_ Dec 7 '17 at 14:32
• yes @CandidMoon and also a tech tree could be "blocked" until a specific tech/skill is acquired or free a good example is Diablo Sklll Tree ( diablo.gamepedia.com/Sorceress_(Diablo_II) ) – Vokail Dec 7 '17 at 15:37