Design
Music
Me
Match Point
TL;DR
Match Point is a tool that helps table tennis players discover equipment tailored to their individual play style, skill level, and goals. Through a short quiz and optional refinement questions, players receive gear recommendations that reflect how they actually play, not just how others review equipment online. Through the use of AI and large language models, Match Point is able to interpret player responses and match them with setups that is personal to them.
By giving players clear, transparent reasons behind each recommendation, Match Point turns a traditionally confusing, subjective process into one that’s simple, personal, and actionable.
Impact at a Glance
High Accuracy: most test cases used in the questionnaire accurately recommended equipment to player archtypes, validated by seasoned professionals and players.
Usage of Recommendations: 5 out of 6 players at the local club who used the tool continue to use setups recommended to them.
Players who have upgraded equipment using this tool have improved their play significantly, citing that their new setup makes them more comfortable.
See final designs
Context & Problem
Personalizing Table Tennis Equipment
Most existing table tennis forums and e-commerce sites are not designed for beginners. Reviews are often written by advanced players, filled with jargon, and lack any personalization. On top of that, numeric ratings (like “Speed: 9.4”) feel arbitrary without meaningful context. As a result, players—especially those still developing their style—end up using gear that doesn’t match how they actually play.
I saw a gap in the market: there was no easy way for players to describe their style and get matched with gear that suits them. My challenge was to build a system that was personal, like a coach who’s seen you play, even if all you gave it was a few sentences and a self-assessment.
My Role
Designer, 2025
Collaboration
Partnered with players and table tennis clubs across the country to validate these recommendations.
Scope
UI Design, Custom ChatGPT Questionnaire
Discovery & Research
Talking to Players and Organizations
To build a product that truly resonated, I talked to many players. I conducted one-on-one interviews with both intermediate and advanced table tennis players, as well as traveling to clubs across the country to observe gameplay styles, ask questions, and better understand the pain points that players experience when choosing equipment. These conversations revealed not only what players were struggling with, but why they were struggling: lack of personalized guidance, unclear gear specs, and conflicting online advice.
Flushing, NY
New Haven, CT
New Orleans, LA
Tampa, FL
Madison, WI
Seattle, WA
In parallel to interviews and field visits, I conducted a competitive analysis of existing table tennis gear recommendation experiences—primarily forums, e-commerce sites, and YouTube reviews.
Across these sources, a few consistent pain points emerged:
Subjectivity
Reviews are highly subjective and often written by players whose needs don't align with beginners or intermediates. They don’t understand a player’s specific needs.
Numeric Ratings
Numeric ratings (e.g. “Speed: 9.4”) lack grounding, making it hard to interpret what a number actually means in practice.
Transparency
No sense of trust or transparency—players don’t understand why something is being recommended to them.
Overabundance
Overwhelming variety of options, with little guidance for someone just getting into custom setups.
Principles
Making Trustworthy Recommendations
Designing Match Point wasn’t just about usability—it was about trust. Through talking with players, I realized that they needed a system that felt personal, reliable, and easy to understand. These principles guided every decision to make sure the product spoke the language of real players.
Transparency
Players should always understand why a setup is recommended to them.
Personalization Without Complexity
Tailor recommendations without overwhelming users with technical jargon or excessive steps.
Confidence Through Context
Frame gear suggestions in player-friendly language and familiar play style concepts, even if they’re new to equipment choices.
Ideation & Exploration
Specific Needs for Specific Play Styles
Generative AI became a compelling solution from the research. If AI could understand a player’s personal playstyle, it would provide massive benefit to equipment recommendations.
I began with a lightweight but targeted questionnaire, intentionally short to reduce friction. The goal was to ask questions that had the highest impact in equipment decisions.
Skill Level
This helps to understand the level of the player taking the questionnaire.
1 of 4
Do you have a USATT rating?
Skill Level Assessment
Yes
No
1 of 4
What is your USATT Rating?
Skill Level Assessment
USATT Rating
Type here
1 of 4
Select the option that best describes your skill level.
Skill Level Assessment
Intermediate
Can rally & developing playstyle
Advanced
Compete regularly, refined technique
Beginner
Still learning the basics
Playstyle Description
This one is the most important and impactful through AI. If a player could accurately describe how they play, we could match keywords to specific pieces of equipment that could tailor their playstyle.
2 of 4
Playstyle Description
Tell us about how you like to play! Be as specific as you can.
Example things to say:
“I like to smash a lot”
“I love to spin the ball”
“I play penhold”
“I do tricky serves”
How do you play?
Type here
Core Shot Ratings
Player confidence in shots also helps to recommend easier or more difficult rubbers and paddles.
3 of 4
Shot Proficiency Ratings
Rate yourself on these shots from a scale of 1 to 5.
1 - Rarely do this, 3 - Can do it inconsistently, 5 - very comfortable
Forehand Loop
Backhand Loop
Serve
Serve Returns
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
Player Intent
This helps to tailor variety and give intentional recommendations.
4 of 4
What are you looking for?
Lastly, what are you looking for in these recommendations?
I want to upgrade my current setup
I’m just curious about what might suit my style
I need a completely new setup that fits my playing style
Complete quiz
Tags
To increase transparency, I produced a set of descriptive tags that formed the backbone of the recommendation system. These tags were surfaced alongside each setup so players could understand exactly why something was recommended based on their answers to the questionnaire.
Using a custom GPT, I gave the questionnaire my specific play style and these were the tags that it produced (which were very accurate):
Intermediate Upgrade
1600 USATT Rating
Defensive Play
Chopping, Long Pips, Modern Defense
Forehand Attack
Offensive Forehand, Not Chopping on FH
Spin-Oriented
Forehand loop-focused attack
Short Game Control
3 in serve/return, critical for modern defense
Testing the Questionnaire
To test the logic, I went to my local club and tested this questionnaire with a custom GPT. I also created dozens of mock player profiles—modern defenders, loopers, penholders, and all-rounders—and manually walked them through the flow. I checked whether the assigned tags aligned with real-world playstyles and known gear pairings, iterating the rules where they didn’t hold up.
This was paired with UX benchmarking of existing forums and shops, which helped guide decisions on tone, layout, and how to expose rubber and blade specs without overwhelming new players.
Final Designs
Welcome to Match Point
The final product is a two-step experience:
Short Questionnaire
Players answer 3–4 high-impact questions that produce immediate, personalized gear recommendations—each with a short explanation and contextual tags.
Recommended Equipment
Includes a blade + rubbers combo, a personalized explanation of the setup and each equipment piece, 2–5 tags for transparency, and multiple variants for different budgets or playstyle learnings.
Tags
Equipment Recommendation
$50
Forehand
Nittaku Hurricane Pro 3 Turbo Blue Sponge
Personalized Explanation
The tone is clear and encouraging, with explanations that feel like advice from a seasoned player or coach.
Reflection & Learnings
Meaningful Features
This project showed me how deeply personal even niche tools can be. For many players, gear is an extension of identity: how they play, how they think, what they aspire to. The key wasn’t giving the most advanced gear, it was giving the gear that made people feel seen.
From a product standpoint, I learned:
- Start simple: earn trust, and invite depth
- Transparency builds confidence: Explaining why matters more than what
- AI can humanize, not just automate: LLMs were crucial in interpreting freeform input and returning smart, scalable matches
Future Plans
In the future, I’d love to expand the product to include three things: conversational follow ups to equipment recommendations, saving equipment recommendations, and a comparison tool for more granularity.
With the help of AI and table tennis becoming more popular, resources and information will become more accessible and can help better inform the system on best recommdnations.
Design
Music
Me
Match Point
TL;DR
Match Point is a tool that helps table tennis players discover equipment tailored to their individual play style, skill level, and goals. Through a short quiz and optional refinement questions, players receive gear recommendations that reflect how they actually play, not just how others review equipment online. Through the use of AI and large language models, Match Point is able to interpret player responses and match them with setups that is personal to them.
By giving players clear, transparent reasons behind each recommendation, Match Point turns a traditionally confusing, subjective process into one that’s simple, personal, and actionable.
Impact at a Glance
High Accuracy: most test cases used in the questionnaire accurately recommended equipment to player archtypes, validated by seasoned professionals and players.
Usage of Recommendations: 5 out of 6 players at the local club who used the tool continue to use setups recommended to them.
Players who have upgraded equipment using this tool have improved their play significantly, citing that their new setup makes them more comfortable.
See final designs
Context & Problem
Personalizing Table Tennis Equipment
Most existing table tennis forums and e-commerce sites are not designed for beginners. Reviews are often written by advanced players, filled with jargon, and lack any personalization. On top of that, numeric ratings (like “Speed: 9.4”) feel arbitrary without meaningful context. As a result, players—especially those still developing their style—end up using gear that doesn’t match how they actually play.
I saw a gap in the market: there was no easy way for players to describe their style and get matched with gear that suits them. My challenge was to build a system that was personal, like a coach who’s seen you play, even if all you gave it was a few sentences and a self-assessment.
My Role
Designer, 2025
Collaboration
Partnered with players and table tennis clubs across the country to validate these recommendations.
Scope
UI Design, Custom ChatGPT Questionnaire
Discovery & Research
Talking to Players and Organizations
To build a product that truly resonated, I talked to many players. I conducted one-on-one interviews with both intermediate and advanced table tennis players, as well as traveling to clubs across the country to observe gameplay styles, ask questions, and better understand the pain points that players experience when choosing equipment. These conversations revealed not only what players were struggling with, but why they were struggling: lack of personalized guidance, unclear gear specs, and conflicting online advice.
Flushing, NY
New Haven, CT
New Orleans, LA
Tampa, FL
Madison, WI
Seattle, WA
In parallel to interviews and field visits, I conducted a competitive analysis of existing table tennis gear recommendation experiences—primarily forums, e-commerce sites, and YouTube reviews.
Across these sources, a few consistent pain points emerged:
Subjectivity
Reviews are highly subjective and often written by players whose needs don't align with beginners or intermediates. They don’t understand a player’s specific needs.
Numeric Ratings
Numeric ratings (e.g. “Speed: 9.4”) lack grounding, making it hard to interpret what a number actually means in practice.
Transparency
No sense of trust or transparency—players don’t understand why something is being recommended to them.
Overabundance
Overwhelming variety of options, with little guidance for someone just getting into custom setups.
Principles
Making Trustworthy Recommendations
Designing Match Point wasn’t just about usability—it was about trust. Through talking with players, I realized that they needed a system that felt personal, reliable, and easy to understand. These principles guided every decision to make sure the product spoke the language of real players.
Transparency
Players should always understand why a setup is recommended to them.
Personalization Without Complexity
Tailor recommendations without overwhelming users with technical jargon or excessive steps.
Confidence Through Context
Frame gear suggestions in player-friendly language and familiar play style concepts, even if they’re new to equipment choices.
Ideation & Exploration
Specific Needs for Specific Play Styles
Generative AI became a compelling solution from the research. If AI could understand a player’s personal playstyle, it would provide massive benefit to equipment recommendations.
I began with a lightweight but targeted questionnaire, intentionally short to reduce friction. The goal was to ask questions that had the highest impact in equipment decisions.
Skill Level
This helps to understand the level of the player taking the questionnaire.
1 of 4
Do you have a USATT rating?
Skill Level Assessment
Yes
No
1 of 4
What is your USATT Rating?
Skill Level Assessment
USATT Rating
Type here
1 of 4
Select the option that best describes your skill level.
Skill Level Assessment
Intermediate
Can rally & developing playstyle
Advanced
Compete regularly, refined technique
Beginner
Still learning the basics
Playstyle Description
This one is the most important and impactful through AI. If a player could accurately describe how they play, we could match keywords to specific pieces of equipment that could tailor their playstyle.
2 of 4
Playstyle Description
Tell us about how you like to play! Be as specific as you can.
Example things to say:
“I like to smash a lot”
“I love to spin the ball”
“I play penhold”
“I do tricky serves”
How do you play?
Type here
Core Shot Ratings
Player confidence in shots also helps to recommend easier or more difficult rubbers and paddles.
3 of 4
Shot Proficiency Ratings
Rate yourself on these shots from a scale of 1 to 5.
1 - Rarely do this, 3 - Can do it inconsistently, 5 - very comfortable
Forehand Loop
Backhand Loop
Serve
Serve Returns
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
Player Intent
This helps to tailor variety and give intentional recommendations.
4 of 4
What are you looking for?
Lastly, what are you looking for in these recommendations?
I want to upgrade my current setup
I’m just curious about what might suit my style
I need a completely new setup that fits my playing style
Complete quiz
Tags
To increase transparency, I produced a set of descriptive tags that formed the backbone of the recommendation system. These tags were surfaced alongside each setup so players could understand exactly why something was recommended based on their answers to the questionnaire.
Using a custom GPT, I gave the questionnaire my specific play style and these were the tags that it produced (which were very accurate):
Intermediate Upgrade
1600 USATT Rating
Defensive Play
Chopping, Long Pips, Modern Defense
Forehand Attack
Offensive Forehand, Not Chopping on FH
Spin-Oriented
Forehand loop-focused attack
Short Game Control
3 in serve/return, critical for modern defense
Testing the Questionnaire
To test the logic, I went to my local club and tested this questionnaire with a custom GPT. I also created dozens of mock player profiles—modern defenders, loopers, penholders, and all-rounders—and manually walked them through the flow. I checked whether the assigned tags aligned with real-world playstyles and known gear pairings, iterating the rules where they didn’t hold up.
This was paired with UX benchmarking of existing forums and shops, which helped guide decisions on tone, layout, and how to expose rubber and blade specs without overwhelming new players.
Final Designs
Welcome to Match Point
The final product is a two-step experience:
Short Questionnaire
Players answer 3–4 high-impact questions that produce immediate, personalized gear recommendations—each with a short explanation and contextual tags.
Recommended Equipment
Includes a blade + rubbers combo, a personalized explanation of the setup and each equipment piece, 2–5 tags for transparency, and multiple variants for different budgets or playstyle learnings.
Tags
Advanced Setup
Two-Wing Offense
Forehand Looping
Spin-Oriented
Stable at Speed
Pressure Game
Equipment Recommendation
$50
Forehand
Nittaku Hurricane Pro 3 Turbo Blue Sponge
Personalized Explanation
The tone is clear and encouraging, with explanations that feel like advice from a seasoned player or coach.
Why this setup works for you
I’d recommend a setup that gives you full control over your pace without sacrificing spin pressure. You're clearly a player who’s comfortable taking initiative from both wings, and your ratings tell me you’ve got the mechanics to handle more responsive, advanced gear.
Based on your responses
See Details
Advanced Setup
Two-Wing Offense
Forehand Looping
Spin-Oriented
Stable at Speed
Pressure Game
Reflection & Learnings
Meaningful Features
This project showed me how deeply personal even niche tools can be. For many players, gear is an extension of identity: how they play, how they think, what they aspire to. The key wasn’t giving the most advanced gear, it was giving the gear that made people feel seen.
From a product standpoint, I learned:
- Start simple: earn trust, and invite depth
- Transparency builds confidence: Explaining why matters more than what
- AI can humanize, not just automate: LLMs were crucial in interpreting freeform input and returning smart, scalable matches
Future Plans
In the future, I’d love to expand the product to include three things: conversational follow ups to equipment recommendations, saving equipment recommendations, and a comparison tool for more granularity.
With the help of AI and table tennis becoming more popular, resources and information will become more accessible and can help better inform the system on best recommdnations.
Design
Music
Me
Match Point
TL;DR
Match Point is a tool that helps table tennis players discover equipment tailored to their individual play style, skill level, and goals. Through a short quiz and optional refinement questions, players receive gear recommendations that reflect how they actually play, not just how others review equipment online. Through the use of AI and large language models, Match Point is able to interpret player responses and match them with setups that is personal to them.
By giving players clear, transparent reasons behind each recommendation, Match Point turns a traditionally confusing, subjective process into one that’s simple, personal, and actionable.
Impact at a Glance
High Accuracy: most test cases used in the questionnaire accurately recommended equipment to player archtypes, validated by seasoned professionals and players.
Usage of Recommendations: 5 out of 6 players at the local club who used the tool continue to use setups recommended to them.
Players who have upgraded equipment using this tool have improved their play significantly, citing that their new setup makes them more comfortable.
See final designs
Context & Problem
Personalizing Table Tennis Equipment
Most existing table tennis forums and e-commerce sites are not designed for beginners. Reviews are often written by advanced players, filled with jargon, and lack any personalization. On top of that, numeric ratings (like “Speed: 9.4”) feel arbitrary without meaningful context. As a result, players—especially those still developing their style—end up using gear that doesn’t match how they actually play.
I saw a gap in the market: there was no easy way for players to describe their style and get matched with gear that suits them. My challenge was to build a system that was personal, like a coach who’s seen you play, even if all you gave it was a few sentences and a self-assessment.
My Role
Designer
2025
Collaboration
Partnered with players and table tennis clubs across the country to validate these recommendations.
Scope
UI Design, Custom ChatGPT Questionnaire
Discovery & Research
Talking to Players and Organizations
To build a product that truly resonated, I talked to many players. I conducted one-on-one interviews with both intermediate and advanced table tennis players, as well as traveling to clubs across the country to observe gameplay styles, ask questions, and better understand the pain points that players experience when choosing equipment. These conversations revealed not only what players were struggling with, but why they were struggling: lack of personalized guidance, unclear gear specs, and conflicting online advice.
Flushing, NY
New Haven, CT
New Orleans, LA
Tampa, FL
Madison, WI
Seattle, WA
In parallel to interviews and field visits, I conducted a competitive analysis of existing table tennis gear recommendation experiences—primarily forums, e-commerce sites, and YouTube reviews.
Across these sources, a few consistent pain points emerged:
Subjectivity
Reviews are highly subjective and often written by players whose needs don't align with beginners or intermediates. They don’t understand a player’s specific needs.
Numeric Ratings
Numeric ratings (e.g. “Speed: 9.4”) lack grounding, making it hard to interpret what a number actually means in practice.
Transparency
No sense of trust or transparency—players don’t understand why something is being recommended to them.
Overabundance
Overwhelming variety of options, with little guidance for someone just getting into custom setups.
Principles
Making Trustworthy Recommendations
Designing Match Point wasn’t just about usability—it was about trust. Through talking with players, I realized that they needed a system that felt personal, reliable, and easy to understand. These principles guided every decision to make sure the product spoke the language of real players.
Transparency
Players should always understand why a setup is recommended to them.
Personalization Without Complexity
Tailor recommendations without overwhelming users with technical jargon or excessive steps.
Confidence Through Context
Frame gear suggestions in player-friendly language and familiar play style concepts, even if they’re new to equipment choices.
Ideation & Exploration
Specific Needs for Specific Play Styles
Generative AI became a compelling solution from the research. If AI could understand a player’s personal playstyle, it would provide massive benefit to equipment recommendations.
I began with a lightweight but targeted questionnaire, intentionally short to reduce friction. The goal was to ask questions that had the highest impact in equipment decisions.
Skill Level
This helps to understand the level of the player taking the questionnaire.
1 of 4
Do you have a USATT rating?
Skill Level Assessment
Yes
No
1 of 4
What is your USATT Rating?
Skill Level Assessment
USATT Rating
Type here
1 of 4
Select the option that best describes your skill level.
Skill Level Assessment
Intermediate
Can rally & developing playstyle
Advanced
Compete regularly, refined technique
Beginner
Still learning the basics
Playstyle Description
This one is the most important and impactful through AI. If a player could accurately describe how they play, we could match keywords to specific pieces of equipment that could tailor their playstyle.
2 of 4
Playstyle Description
Tell us about how you like to play! Be as specific as you can.
Example things to say:
“I like to smash a lot”
“I love to spin the ball”
“I play penhold”
“I do tricky serves”
How do you play?
Type here
Core Shot Ratings
Player confidence in shots also helps to recommend easier or more difficult rubbers and paddles.
3 of 4
Shot Proficiency Ratings
Rate yourself on these shots from a scale of 1 to 5.
1 - Rarely do this, 3 - Can do it inconsistently, 5 - very comfortable
Forehand Loop
Backhand Loop
Serve
Serve Returns
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
Player Intent
This helps to tailor variety and give intentional recommendations.
4 of 4
What are you looking for?
Lastly, what are you looking for in these recommendations?
I want to upgrade my current setup
I’m just curious about what might suit my style
I need a completely new setup that fits my playing style
Complete quiz
Tags
To increase transparency, I produced a set of descriptive tags that formed the backbone of the recommendation system. These tags were surfaced alongside each setup so players could understand exactly why something was recommended based on their answers to the questionnaire.
Using a custom GPT, I gave the questionnaire my specific play style and these were the tags that it produced (which were very accurate):
Intermediate Upgrade
1600 USATT Rating
Defensive Play
Chopping, Long Pips, Modern Defense
Forehand Attack
Offensive Forehand, Not Chopping on FH
Spin-Oriented
Forehand loop-focused attack
Short Game Control
3 in serve/return, critical for modern defense
Testing the Questionnaire
To test the logic, I went to my local club and tested this questionnaire with a custom GPT. I also created dozens of mock player profiles—modern defenders, loopers, penholders, and all-rounders—and manually walked them through the flow. I checked whether the assigned tags aligned with real-world playstyles and known gear pairings, iterating the rules where they didn’t hold up.
This was paired with UX benchmarking of existing forums and shops, which helped guide decisions on tone, layout, and how to expose rubber and blade specs without overwhelming new players.
Final Designs
Welcome to Match Point
The final product is a two-step experience:
Short Questionnaire
Players answer 3–4 high-impact questions that produce immediate, personalized gear recommendations—each with a short explanation and contextual tags.
Recommended Equipment
Includes a blade + rubbers combo, a personalized explanation of the setup and each equipment piece, 2–5 tags for transparency, and multiple variants for different budgets or playstyle learnings.
Tags
Advanced Setup
Two-Wing Offense
Forehand Looping
Spin-Oriented
Stable at Speed
Pressure Game
Equipment Recommendation
$50
Forehand
Nittaku Hurricane Pro 3 Turbo Blue Sponge
Personalized Explanation
The tone is clear and encouraging, with explanations that feel like advice from a seasoned player or coach.
Why this setup works for you
I’d recommend a setup that gives you full control over your pace without sacrificing spin pressure. You're clearly a player who’s comfortable taking initiative from both wings, and your ratings tell me you’ve got the mechanics to handle more responsive, advanced gear.
Based on your responses
See Details
Advanced Setup
Two-Wing Offense
Forehand Looping
Spin-Oriented
Stable at Speed
Pressure Game
Reflection & Learnings
Meaningful Features
This project showed me how deeply personal even niche tools can be. For many players, gear is an extension of identity: how they play, how they think, what they aspire to. The key wasn’t giving the most advanced gear, it was giving the gear that made people feel seen.
From a product standpoint, I learned:
- Start simple: earn trust, and invite depth
- Transparency builds confidence: Explaining why matters more than what
- AI can humanize, not just automate: LLMs were crucial in interpreting freeform input and returning smart, scalable matches
Future Plans
In the future, I’d love to expand the product to include three things: conversational follow ups to equipment recommendations, saving equipment recommendations, and a comparison tool for more granularity.
With the help of AI and table tennis becoming more popular, resources and information will become more accessible and can help better inform the system on best recommdnations.