How Talent in Text Adventures Affects the Same Action: Change the Information, Not Just Add a Compliment
The same action, with different talents, should produce different information and different available paths, rather than the same result paired with three compliments. The method is: first fix the action itself, then for each talent write one "clue only it can see" and one "follow-up only it can take," and finally add an alternative path for players without that talent. Below, a fictional teaching example, "Night Rain Train Signal Failure," breaks it down. Character names, numbers, and dialogue are all teaching setups, not tested data.
Introduction
The same action, with different talents, should produce different information and different available paths, rather than the same result paired with three compliments. The method is: first fix the action itself, then for each talent write one "clue only it can see" and one "follow-up only it can take," and finally add an alternative path for players without that talent. Below, a fictional teaching example, "Night Rain Train Signal Failure," breaks it down. Character names, numbers, and dialogue are all teaching setups, not tested data.
First Fix the Action, Then Split by Talent
Suppose the player is in the third carriage of the fictional "Night Rain" train, and the action is "inspect the signal box." This action is the same sentence for all three talents, but the output should be divided into three layers:
| Layer | Function | Example for inspecting the signal box |
|---|---|---|
| Observation layer | Gives information only that talent can notice | Investigation: the relay contacts have fresh scratches; Repair: the coil temperature is abnormal; Negotiation: the duty officer avoids mentioning "last night's shift change" |
| Path layer | Gives one follow-up only that talent can take | Investigation: can press about the source of the scratches; Repair: can try a short-circuit test; Negotiation: can ask the duty officer to swap shift records |
| State layer | Records what this action changed | Records "signal box opened," "duty officer becomes guarded," etc., for later reference |
The key is the observation layer. If all three talents only get "the signal box is broken," then the difference is just one compliment, and the player cannot feel the talent doing anything. The observation layer must give a concrete thing that can be followed up on: scratches, temperature, an avoided topic—all are hooks for later actions. The state layer must be restrained, recording only what this action truly changed, and not writing in things the player did not do.
Complete Writing for Three Types of Talent Turns
The following three passages are three fictional demonstrations of the same action.
Investigation talent. The player says "inspect the signal box." Output: "You crouch down, flashlight sweeping across the inside of the box. There are three parallel scratches on the relay contacts, their edges still bright, as if just scraped by something hard. The duty officer stands behind you, saying nothing." Options: press about the scratches, first note the position, pretend not to see. This scratch is information unique to the investigation talent; other talents cannot see it. It does not directly say who made it, only gives the player a direction to keep pursuing.
Repair talent. The same sentence, "inspect the signal box." Output: "You put the back of your hand close to the coil; it is hotter than normal. A spot of insulation has gone soft. You judge it will not last until the next track change." Options: short-circuit test, remove the spare coil, cut power before checking. Temperature and softness are information unique to the repair talent. It does not guarantee a successful repair, only tells the player where the fault may be and what risks acting now carries.
Negotiation talent. The same sentence, "inspect the signal box." Output: "You ask the duty officer when the last maintenance was. He says, 'According to the schedule,' but his eyes look out the window. You notice that when he mentions 'last night's shift change,' his speech speeds up by a beat." Options: press about the shift change, chat about something else first and then circle back, directly demand to see the records. Avoidance and speech speed are information unique to the negotiation talent. It does not prove the other person is lying, only gives the player a doubt worth pressing on.
In all three passages, the action is the same, the information is different, and the follow-up paths are also different. None of the talents directly gives "repaired" or "asked it out"—they only change what the player can see and where the player can go. Whether the player presses, short-circuits, or applies pressure is still decided by later choices.
Alternative Paths Without That Talent
If the player does not have the investigation talent, the scratch information does not appear, but the action must not become a dead end. The alternative is to give a clumsier, slower, more other-dependent path: the player can "ask the duty officer to look together," and the other party points out the contact position, but the other party may misjudge or conceal it. In this way, players without the talent can still make progress, but the information quality and cost differ.
Without the repair talent, the player can "check item by item according to the manual," which takes longer, and the manual may not cover the current fault; the player can also "call a maintenance worker," but must wait, and after the other party arrives, the scene may change. Without the negotiation talent, the player can "directly flip through the duty records," but may be stopped or misremember, and may also find a form that has been altered.
Alternative paths must clearly state the cost, rather than letting players without the talent automatically succeed. The cost can be time, another person's involvement, reduced credibility of information, or worsening some state. When writing, ask: what does this alternative path make the player lose? If nothing is lost, it will become a more cost-effective choice than the talent path, and the talent difference is erased again.
Avoid a Single Mandatory Choice
If all three talents point to the same "correct answer," players will still feel there is only one path. The way to check is: put the three outputs side by side and ask, "If the player only takes one of them, will they miss the key information of the other two?" If yes, the information is complementary, not repetitive. Then ask, "Is any path obviously shorter and safer?" If so, give it a cost, or give the other paths a unique reward.
You can also do a small comparison: cover the investigation talent's scratches, the repair talent's softness, and the negotiation talent's speech speeding up by a beat, and see whether the remaining two can still progress independently. If covering one immediately makes the other two stall, then this information has been written as a mandatory checkpoint, and it needs to be downgraded to "faster or lower cost," rather than "cannot get through without it."
Completion Check
After writing the three types of talent turns for one action, check them one by one:
- Is the action sentence exactly the same?
- Does each talent have one piece of information only it can see?
- Does each talent have one follow-up only it can take?
- Without that talent, is there an alternative path with a cost?
- Are the three paths complementary, rather than pointing to the same answer?
- Does the state layer record only what this action truly changed?
If all six pass, then this action truly lets talent change the information, rather than just adding a compliment. Next, you can pick one action in your script that players repeat most often, rewrite its three types of turns according to this table, and cover one of them to check whether the other two can still progress independently.
