💻 computer science

When Friction Helps: Transaction Confirmation Improves Decision Quality in Blockchain Interactions

Contrary to the prevailing view that transaction confirmations are merely usability friction to be minimized, this study demonstrates that requiring manual confirmation in blockchain interactions actually improves decision quality and enables self-correction, despite users' preference for frictionless, auto-authorized flows.

Original authors: Eason Chen, Xinyi Tang, George Digkas, Dionysios Lougaris, John E. Naulty, Kostas Chalkias

Published 2026-02-24

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Eason Chen, Xinyi Tang, George Digkas, Dionysios Lougaris, John E. Naulty, Kostas Chalkias

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Idea: Sometimes, a Speed Bump is a Good Thing

Imagine you are driving a car. You want to get to your destination as fast and smoothly as possible. You hate hitting potholes or stopping at red lights. In the world of technology, especially blockchain (like crypto apps), designers have been trying to remove all the "speed bumps." They want you to click a button and have your transaction happen instantly, without any annoying pop-ups asking, "Are you sure?"

This paper asks a simple question: What if those annoying pop-ups (the "friction") are actually saving us from making stupid mistakes?

The researchers found that while users love the smooth, fast way of doing things, they actually make better decisions when they are forced to pause and confirm their actions.

The Experiment: A Digital Game of Connect Four

To test this, the researchers built a version of the classic game Connect Four that runs on a blockchain. They set up two different ways to play:

The "Smooth" Mode (Frictionless): You click a square to drop your piece, and it happens instantly. No questions asked. It feels like playing a normal video game.
The "Bumpy" Mode (Confirmation): You click a square, but before your piece drops, a little window pops up (like a wallet confirmation) asking, "Are you sure you want to move here?" You have to click "Confirm" to proceed.

The Setup:
They had over 100 people play this game. Some played in "Smooth" mode, some in "Bumpy" mode, and some played both. The game was played against a smart computer opponent.

The Surprising Results

Here is where it gets interesting. The results showed a massive split between what people felt and what they actually did.

1. The Feeling: "I'm a Pro!"

When asked after the game, everyone said the Smooth Mode was better.

They felt less frustrated.
They felt they were playing faster.
They felt like they made fewer mistakes.
Analogy: It felt like driving a Ferrari on an open highway. Everyone loved the speed.

2. The Reality: "Oops, I Lost."

When the researchers looked at the actual game scores, the Smooth Mode players were losing more often.

Their win rate dropped by nearly 12%.
Their moves were objectively worse (the computer rated their strategy lower).
Analogy: Even though they felt like they were driving a Ferrari, they were actually driving off a cliff because they were going too fast to see the road.

The Lesson: Removing the "stop and think" moment made people feel confident, but it made them careless.

The "Second Chance" Mechanism

The researchers dug deeper to see why the "Bumpy" mode worked better. They looked at the moments when players in the "Bumpy" mode clicked a move, saw the confirmation pop-up, and then changed their mind.

The Scenario: A player clicks a square, the pop-up appears, and they think, "Wait, if I put my piece there, my opponent will win next turn!" They hit "Cancel," pick a different square, and confirm that one instead.
The Result: When players did this "self-correction," their new move was almost always better than the first one they picked.
Analogy: It's like writing a text message, hitting send, and then realizing you made a typo. The "confirmation pop-up" is like that split second where you see the message before it goes out, giving you a chance to fix it. Without that pause, the typo gets sent forever.

Why Does This Matter?

In the world of Web3 and crypto, companies are racing to make apps "frictionless" (no pop-ups, no confirmations) to make them easier to use. They think friction is the enemy.

This paper argues that friction is actually a friend.

The Trade-off: If you remove the friction, you get a smoother experience, but you lose decision quality.
The "Commitment Boundary": The confirmation pop-up acts as a "stop sign" or a "checkpoint." It forces your brain to switch from "autopilot" mode to "thinking" mode. It reminds you: "Hey, this action is permanent. You can't undo it. Are you really sure?"

The Takeaway for Everyday Life

Think of it like signing a contract or buying something expensive online.

Frictionless: You swipe your card, and the money is gone instantly. You feel great because it was easy. But later, you realize you bought something you didn't need.
With Friction: You have to type in a password or confirm the amount. It feels annoying for a second. But that tiny annoyance gives your brain a moment to say, "Wait, do I really need this?"

Conclusion:
The authors suggest that we shouldn't just try to remove all the "speed bumps" in technology. Instead, we should design systems that keep the ride smooth but add smart speed bumps at the most critical moments. These "responsible frictions" act as a safety net, helping us think before we act, ensuring that when we make a permanent decision, we are actually making the right one.

1. Problem Statement

In the Web3 ecosystem, a dominant design trend aims to minimize "friction" by automating transaction authorization (e.g., using session keys, embedded wallets, or zero-knowledge logins) to mimic the seamless experience of traditional web applications. While this reduces usability barriers, the authors hypothesize that removing explicit transaction confirmation steps may inadvertently degrade human decision-making quality.

The core tension lies between usability optimization (smooth, automated flows) and decision quality in irreversible contexts. Classic security-usability research suggests that when interfaces conflict with mental models, errors occur. However, this study investigates the counter-intuitive possibility that the "friction" of a confirmation dialog serves a functional cognitive role: acting as a checkpoint for deliberation, verification, and self-correction before an irreversible commitment.

2. Methodology

The authors conducted a controlled user study using a blockchain-based Connect Four game built on the Sui network. The study isolated the variable of "authorization flow" while keeping all other factors (game rules, UI, opponent AI) constant.

Experimental Design: A within-subject design where participants played in two distinct modes:
1. Confirmation Mode: Every move required manual wallet confirmation via a browser extension pop-up.
2. Frictionless Mode: Moves were automatically authorized using a pre-established session key or ephemeral private key, eliminating the pop-up.
Participants: 109 eligible participants (recruited from X/Twitter blockchain communities with prior wallet experience) across two deployment waves.
- Wave 1: Randomized per game (no counter-balancing).
- Wave 2: Strictly counter-balanced (ensuring equal exposure to both modes in the first four games).
AI Engine: A hybrid engine using Monte Carlo Tree Search (MCTS) for adaptive opponent behavior and Minimax search (depth 4) for deterministic evaluation of board states. This allowed for the calculation of an objective "move quality" score (0 to 1) for every player action relative to the optimal move.
Metrics:
- Subjective: Post-test surveys (7-point Likert scale) measuring perceived usability, attention switching, frustration, and self-assessed performance.
- Objective: Win rate and AI-calculated move quality scores.
- Behavioral: Analysis of "rejected moves" (cases where a user opened the confirmation dialog, canceled the transaction, and selected a different move).

3. Key Contributions

The paper makes two primary contributions to the fields of Usable Security and Human-Computer Interaction (HCI):

Dissociation of Perceived vs. Objective Performance: It demonstrates that while users prefer and perceive higher performance in frictionless (automated) modes, their actual objective decision quality is significantly lower compared to modes with manual confirmation.
Confirmation as a Cognitive Checkpoint: It provides empirical evidence that transaction confirmation dialogs function as "reflective checkpoints." They enable pre-submission self-correction, allowing users to reconsider and improve their decisions before an irreversible action is committed.

4. Key Results

The study yielded statistically significant findings highlighting a trade-off between interaction smoothness and decision quality:

Subjective Preference: Participants strongly preferred the Frictionless Mode. They reported higher perceived performance, less frustration, and fewer attention switches compared to the Confirmation Mode (all $p < .001$ ).
Objective Performance Decline: Despite the preference for frictionless interaction, removing confirmation led to worse outcomes in the counter-balanced Wave 2 data:
- Win Rate: Dropped by 11.8% in Frictionless Mode (9.5%) vs. Confirmation Mode (21.2%) ( $p = 0.044$ ).
- Move Quality: The average quality of moves decreased by 0.051 in Frictionless Mode ( $p = 0.022$ ).
Self-Correction Mechanism: Analysis of rejected moves in the Confirmation Mode showed that when users canceled a transaction, their subsequent "correction" move was significantly better.
- Original rejected move quality: 0.616.
- Submitted correction move quality: 0.737.
- Improvement: +0.121 ( $p = 0.0052$ ).
- In 45.5% of cancellation cases, the user successfully selected a higher-quality move after reconsidering.

5. Significance and Implications

The findings challenge the prevailing Web3 design dogma that "friction is always bad."

Responsible Friction: The authors argue for "responsible friction"—designs that maintain smooth flows but reintroduce lightweight verification opportunities at critical commitment boundaries.
Calibrated Confidence: There is a "responsibility gap" where users feel more confident in automated systems but make more errors. Future systems need to align user confidence with actual decision quality.
Beyond Binary Design: The study suggests moving away from a binary choice (fully manual vs. fully automated) toward a spectrum of confirmation designs. For example, low-risk actions could be automated, while high-risk or unusual actions should trigger explicit confirmation to prompt deliberation.
Cognitive Role of Interruption: Unlike external interruptions that degrade performance, the authors posit that self-triggered confirmation dialogs serve as a psychological "reset," breaking autopilot patterns and signaling the irreversibility of the action, thereby improving decision quality in irreversible contexts (trading, gaming, legal contracts).

Conclusion: While automation improves the experience of using blockchain applications, removing transaction confirmation can systematically degrade the quality of decisions. Effective blockchain UX design must balance usability with mechanisms that support human deliberation and self-correction.