📈 economics

Pairwise Beats All-at-Once: Behavioral Gains from Sequential Choice Presentation

This paper proposes and tests the Sequential Rationality Hypothesis, arguing that presenting products through sequential pairwise comparisons rather than simultaneous all-at-once displays reduces cognitive overload and context-dependent errors, thereby enabling consumers to make more utility-maximizing and rational decisions.

Original authors: Dipankar Das

Published 2026-01-23

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Dipankar Das

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

Imagine you are walking into a massive, chaotic supermarket to buy a drink. The shelves are packed with hundreds of options: Coffee, Tea, Juice, Soda, Water, and dozens of exotic blends. All of them are staring at you at once, with flashing "Sale!" signs, star ratings, and colorful labels.

This is how most digital markets (like Amazon or Netflix) work today: The "All-at-Once" Approach.

The paper argues that when you are bombarded with too many choices at the same time, your brain gets overwhelmed. You might miss the best option entirely because you're too busy trying to process the noise. You might pick a "good enough" drink just to stop the stress, even if there was a perfect one right next to it that you didn't notice.

The New Idea: "Pairwise Beats All-at-Once"

The author, Dipankar Das, proposes a different way to shop, which he calls the Sequential Rationality Hypothesis.

Instead of showing you the whole shelf, imagine a smart shopkeeper who only shows you two drinks at a time.

Step 1: The shopkeeper puts a Coffee and a Tea in front of you. "Which one do you prefer?" You pick the Coffee.
Step 2: The shopkeeper takes the Coffee and puts it next to a Juice. "Now, which of these two do you want?" You pick the Coffee again.
Result: You walk out with the Coffee.

The paper claims that this "one-on-one" tournament style helps you make a better, more rational decision than staring at a giant list of ten drinks.

Why Does This Work? (The "Brain Battery" Analogy)

Think of your attention like a battery.

The All-at-Once Method: Trying to compare 10 items at once drains your battery instantly. Your brain gets tired (cognitive overload), and you start making mistakes. You might pick the drink with the biggest "Sale" sticker instead of the one that actually tastes best.
The Sequential Method: Comparing just two items uses very little battery. You can focus 100% on the difference between those two. By the time you move to the next pair, your battery is still fresh.

The paper uses a mathematical model (called the "Random Attention Model") to prove that if you break a big choice down into small, two-item battles, you are statistically more likely to end up with the item you truly love, rather than the one that just looked loudest in the crowd.

The "Tournament" Metaphor

Think of it like a sports tournament.

Simultaneous Choice: Imagine a referee trying to pick the best player from a crowd of 100 people all shouting at once. The referee gets confused and picks someone who is just loud, not necessarily the best.
Sequential Choice: Imagine a knockout tournament. Player A fights Player B. The winner fights Player C. The winner fights Player D.
- In every single match, the referee only has to watch two people. It's easy to see who is better.
- By the end of the tournament, the winner is almost certainly the best player, because they had to beat everyone else in fair, focused one-on-one fights.

What Does This Mean for You?

The paper suggests that digital platforms (apps, websites) are currently designed to sell you things by overwhelming you with choices and distractions. This often leads to "irrational" choices where you buy something you don't really want.

The author argues that if these platforms changed their design to show you two options at a time (like a "this or that" slider), you would:

Feel less stressed.
Make fewer mistakes.
Be more likely to buy the product that actually gives you the most happiness.

The Catch (Limitations)

The paper admits that this idea is currently a theory. The author has done the math and the logic proofs (even checking them with computer software called Lean4 to be sure), but they haven't run a real-world experiment with actual people yet because they don't have the funding or data for that right now.

In short: The paper claims that if you stop trying to choose from a mountain of options and start choosing from a series of simple, two-item battles, you will make smarter, happier decisions. It's a call for digital stores to stop shouting at you and start whispering to you, one pair at a time.

Technical Summary: Pairwise Beats All-at-Once: Behavioral Gains from Sequential Choice Presentation

Problem Statement
The paper addresses a fundamental disconnect in digital markets between classical economic assumptions of full rationality and the empirical reality of bounded rationality. While classical models assume agents can process and evaluate all alternatives simultaneously to maximize utility, digital interfaces often present diverse product assortments simultaneously (e.g., grid layouts on Amazon). This "all-at-once" presentation frequently leads to cognitive overload, attention fragmentation, and suboptimal choices due to the limitations of human cognitive capacity. The paper posits that current choice architectures, which rely on simultaneous multi-option displays, fail to align with cognitive processing capabilities, leading to inconsistent and irrational consumer decisions.

Methodology
The authors develop a formal theoretical framework grounded in the Random Attention Model (RAM) (Cattaneo et al., 2020) to model decision-making under bounded attention. The core methodology involves:

Theoretical Modeling: The paper constructs a "Sequential Rationality Hypothesis" (SRH) which contrasts two choice architectures:
- Simultaneous Offering: The decision-maker (DM) evaluates a set $S$ of $n$ items at once.
- Sequential Pairwise Offering: The DM evaluates items through a recursive binary tournament (e.g., comparing $A$ vs. $B$ , then the winner vs. $C$ ).
Formal Definitions: The model utilizes a strict preference ordering ( $\succ$ ) and a latent attention rule $\mu(T|S)$ , representing the probability that a consideration set $T \subseteq S$ is formed. A key axiom introduced is Monotonic Attention, which posits that removing non-considered alternatives from a set weakly increases the probability of attending to a specific subset.
Axiomatic Characterization: The authors define a choice rule as sequentially rational if it satisfies specific axioms: Binary Consistency, Sequential Transitivity, Attention Monotonicity, and Cognitive Parsimony.
Probabilistic Analysis: The paper derives probability inequalities comparing the likelihood of selecting the utility-maximizing option under both architectures. It introduces parameters for attention success ( $\alpha_k$ ) and choice accuracy ( $\beta_k$ ) dependent on the arity (number of items) of the comparison.
Numerical Illustration: A concrete example using three beverages (Coffee, Tea, Juice) with a specified monotonic attention rule is used to compute and compare choice probabilities under RAM for both simultaneous and sequential scenarios.
Formal Verification: All mathematical definitions, theorems, and proofs are formally verified using the Lean4 proof assistant, ensuring logical rigor.

Key Contributions

Sequential Rationality Hypothesis (SRH): The paper introduces a new behavioral mechanism arguing that sequential, pairwise presentation of options systematically enhances decision quality by aligning choice structures with cognitive constraints.
Extension of RAM: The work extends the Random Attention Model by integrating a recursive, dynamic decision process. Unlike static list-based models or network traversal models, SRH explicitly models how attention constraints interact with sequential binary comparisons to shape decision paths.
Theoretical Proof of Superiority: The authors prove that under bounded attention, the probability of selecting the utility-maximizing option is strictly higher in a sequential binary tournament than in a simultaneous evaluation, provided that the effective attention accuracy in binary comparisons is sufficiently high relative to ternary (or larger) comparisons.
Cognitive Cost Reduction: The paper demonstrates mathematically that decomposing a choice task into binary steps reduces total expected cognitive costs compared to evaluating a large set simultaneously, leveraging the convexity of attention cost functions.

Results

Probability Inequalities: The central result (Theorem 3 and Theorem 4) establishes that:
$Pr(C_{SEQ}(A, B, D) = \text{arg max } u) > Pr(C_{SIM}(\{A, B, D\}) = \text{arg max } u)$
This holds if the probability of a correct choice in a binary comparison ( $q$ ) exceeds the square root of the probability of a correct choice in a ternary comparison ( $r$ ), i.e., $q > r^{1/2}$ .
Numerical Validation: In the provided numerical example, the probability of choosing the optimal item (Coffee) was 0.81 under the sequential architecture versus 0.60 under the simultaneous architecture, quantifying the behavioral gain.
Divergence from Standard Theory: The model shows that $C_{SEQ}(S) \neq C_{SIM}(S)$ when attention is limited, providing a theoretical justification for why standard revealed preference theory (which assumes full attention) fails to predict choices in complex digital environments.
Equivalence Conditions: The paper proves that sequential and simultaneous choice operators coincide only under conditions of full attention ( $\mu(S|S)=1$ ) and deterministic maximization, conditions rarely met in real-world digital markets.

Significance and Claims
The paper claims to bridge the gap between cognitive limitations and economic rationality by embedding cognitive realism into decision theory. Its significance lies in:

Platform Design: It offers a theoretical basis for digital platforms to redesign recommendation flows from grid-based, simultaneous displays to sequential, pairwise comparisons to reduce cognitive overload and improve consumer welfare.
Policy Implications: It suggests that policymakers and regulators in digital markets (where specific laws are currently lacking) could incentivize "attention-friendly" interfaces that structure choices sequentially to prevent context-dependent errors and irrational decisions.
Theoretical Advancement: It provides a middle ground between full rationality and unstructured heuristics, offering a formal, falsifiable framework for understanding how choice architecture influences rationality under bounded attention.

The authors explicitly note that the current work is theoretical due to a lack of suitable empirical data, and they identify the need for future experimental designs to test these hypotheses empirically. The paper does not claim to have solved the problem of digital market regulation but rather provides the theoretical underpinning for how choice architecture can be optimized.