Level Design Based on the Number of Failures
Clear rate VS Number of failures
In the meantime, when discussing the difficulty of the puzzle level, we have designed the difficulty and measured the data based on the clear rate, which is how much the level has been cleared. Because it uses popular probabilities, it has the advantage of being easy for anyone to calculate and handle, but it has the disadvantage that it is completely different from the actual number and the play experience that the user experiences directly.
(image)
Correlation graph of clearing rate by number of failures
Usually, when a user plays twice and fails once, the clear rate is measured to be about 50%. Similarly, if you fail four times in five plays, the clearance rate is about 20%, and if you fail nine times in ten plays, the clearance rate is measured to be about 10%. As the clearance rate decreases, the number of probabilities changes little by little, while the number of failures increases exponentially.
The harder the level, the larger the difference of 1% or less, and the harder it is to make accurate predictions with only probabilities when we want to induce an appropriate failure experience for the user. Therefore, I would like to take the actual number of failures of users as a simple and clear standard.
Appropriate number of failures
Then, users can think about how much failure they experience and how much they will have a desire to win at the level and continue to challenge in a fun way. You can also think about how you can continue to learn and grow when each user has different skills and tendencies. In order to solve the above concerns, the number of failures by level section of reference games with high retention and good performance among in-house games was measured and listed as a distribution map.
(image)
Distribution chart of the number of level failures of two in-house games
In the early level section of the game, the distribution of small failures was very easy from the perspective of learning and adapting users, and the difficulty was gradually increased in the middle level section, and the experience was increased and maintained above a certain level in the second half. In particular, the second-half level section moves only within the appropriate number of failures, and it is possible to assume that users will continue to challenge with the greatest fun when they do not bounce or fall more or less.
Therefore, the proper failure experience brings great benefits to users and games, and even if it is difficult to accurately grasp, the key is to find the appropriate number of failures to simultaneously capture the user’s fun and the game’s performance. When designing a puzzle level based on it, it will be best for everyone to stimulate a sense of challenge by properly mixing easy and difficult levels, giving users sufficient learning and experience, and then applying previously learned things in earnest.
How to design level difficulty
As mentioned in the distribution diagram, since the level must be designed within the appropriate number of failures, it is recommended that the width be determined according to the level section when setting the level difficulty. The initial level section needs to be designed with detailed controls by narrowing the width with a small number of failures, and at the mid to late level, it needs to be designed flexibly by providing users with whip and carrot at the same time.
However, neither too powerful a whip nor too sweet a carrot can cause users great pain and boredom. A level that is too difficult will even dampen the sense of challenge and cause users to feel frustrated, while a level that is too easy will cause users to forget the reason to play the game and lose their fun.
(image)
Flexible design of failure range as target, but too high difficulty becomes toxic
Therefore, a Maginot Line of difficulty is always required, and the level should be designed to be successful by utilizing the experience when the user has experienced enough failures. If you experience continuous failure, you challenge yourself steadily and repeatedly without the knowledge you gain from itEven so, the user is bound to be unable to move forward any more. To prevent it in advance, difficult levels beyond a certain number of failures should be truncated so that users can approach success more closely. (Cut-off)
In this way, after setting the range of the number of failures for each level section and setting the maximum number of failures to the cut-off standard, a level difficulty design plan, which is the start of the level design task, is completed. Then, if you look at the blueprint, produce an interesting level, analyze the user’s response again, modify the blueprint, or modify the level to continuously improve it, you will surely be able to provide the fun of the puzzle that the user wanted.