Writing · Free lesson
“Better” on what? Compare model results precisely
Two models were checked on the same 100 labeled examples. Model A had 89% accuracy; Model B had 92%. What can your report actually claim?
Professor Scott MendozaThe short answer
Instead of “Model B is better,” say what was higher or lower and where it was measured: “On evaluation set A, Model B had higher accuracy than Model A.”
Keep the comparison attached to its measurement.
The result supports higher accuracy for B on that set. It does not establish that B is faster, cheaper or better for every user. “On this evaluation set” limits the claim to the evidence you have. Use a different metric name when you are discussing a different measurement.
A comparison with a clear scope
On evaluation set A, Model B had 92% accuracy, compared with 89% for Model A.
This result does not tell us which model is faster.
Higher accuracy and lower response time describe different measurements. Do not silently swap one for the other.
Include the missing context
Name the set
On the same 100 labeled examples,
Name the metric and comparison
Model B had higher accuracy than Model A.
Try it now
Keep the comparison within the evidence
Read all answers and explanations
Same set: A accuracy 89%; B accuracy 92%. Which statement is supported?
B had higher accuracy on this set.The stated measurement is accuracy on one set. Speed and all future tasks were not measured.
A has 90% accuracy on set X. B has 93% on a different set Y. Which report keeps that difference visible?
A scored 90% on X; B scored 93% on Y. These are different evaluation sets.Report the two results with their set names. Different sets should not disappear from the comparison.
The only measurement is response time: A 200 ms, B 150 ms. Which sentence fits?
B had a lower response time in this measurement.150 ms is lower than 200 ms. The measurement is response time, not accuracy.
Put it into practice
Repair a broad experiment claim
On evaluation set C, Model A had 81% accuracy and Model B had 84%. No speed, cost or production result is given. Choose a precise summary.
Model B is better in every way and ready for production.
See a model answer
On evaluation set C, Model B had 84% accuracy, compared with 81% for Model A.
Other wording is valid if it keeps the same metric, values and evaluation context.
A little more clarity
Common questions
A few useful details before you move on.
Can I use outperformed?
Yes, but state the measure and context. “Outperformed on accuracy on this set” is narrower than “outperformed in every way.”
Model B outperformed Model A on accuracy on evaluation set C.
How can I report equal results on one metric?
Say that the models achieved the same result on that metric and evaluation set. This does not mean the models are identical or equally good in every respect.
Take it into your day
A precise comparison is useful because its limits are visible.
English for Careers · Coming Soon
Present experiment results in team English.
English for Machine Learning Engineers: Career Simulation is coming soon, with planned scenarios involving experiments, evaluation and model monitoring.
Get launch updatesMachine Learning Engineers · Career Simulation
Your next step · Available on Udemy
Build clearer sentences for work.
The available English Grammar and English Speaking Master Course develops sentence-building, vocabulary and business-communication foundations. It is a broader English course, not this specialist career simulation.
Explore the course English Master Course: English Grammar and English Speaking