HireVue Games and Assessment Practice
HireVue sells four different things through one platform. This page is the assessment side: the eleven games, the Virtual Job Tryout and CodeVue, built to the scoring model HireVue actually publishes.
Free Tests
Free to try with a free account - unlock this provider for unlimited practice.
Premium Tests
Singularity
Find the one shape that differs, as distractors multiply and the difference narrows to a single feature.
Pathfinder
Slide the tiles to build a continuous path between two endpoints, with five minutes on the clock.
Digitspan
Reproduce alphanumeric sequences that lengthen as you climb, with later levels asking for them reordered.
Flashback
Say whether the shape on screen matches one from n steps back, with n rising level by level.
Puzzle Picture
Memorise an image, then rebuild it from scrambled pieces inside seven minutes.
E-Motions
Identify the emotion on each face against a four-minute clock. An ability test despite the name: HireVue validates it against a performance measure.
Portrait
Choose between paired images to describe your own working style, about fifty times. No right answers.
PortraitXT
Rate your agreement with about fifty statements about how you work. No right answers.
Teamchat
Reply to a branching team conversation in a chat thread, three options a turn. Describes how you work with people.
Virtual Job Tryout
A multi-section job simulation: situational judgement, a multi-source information task, emotion recognition, a delayed recall trap and a work-style section.
CodeVue
A coding challenge with automatically scored test cases. Runs JavaScript here; the real one uses your employer’s language.
Games or interview? Check the invitation before you practise
Verified September 2026You are in the right place if your invitation mentions a game-based assessment, or names any of Numerosity, Shapedance, Singularity, Pathfinder, Digitspan, Flashback, Puzzle, E-Motions, Portrait, PortraitXT or Teamchat, or a Virtual Job Tryout or CodeVue coding assessment.
If it only names a recorded video interview, an OnDemand Interview, or asks you to record answers to camera, the practice you want is our interview practice, not this page.
It is very often both.HireVue's own white paper describes a graduate assessment as six interview questions plus three short game challenges in one sitting, so an email mentioning a video interview does not rule out games. The two are marked separately.
HireVue no longer analyses facial expressions. It discontinued the feature in March 2020 and announced that in January 2021. Candidates still arrive believing their face is being scored and spend the interview managing their expression instead of answering.
The level ladder, which is the whole mechanic
Up two, down one
HireVue describes it in its own white paper: each level increases in difficulty, and players move to a higher level each time they complete one successfully, or down if they fail. Difficulty rises through added shapes, colours and distraction effects. The cap is level 100 for every game, and nobody gets near it: the observed mean final level was 39 on Numerosity and 27 on Shapedance.
One clock per game, not per question
Each game runs about three minutes on its own clock, shown next to the level number. There is no per-item timer and no overall battery timer. The game ends when the clock ends, never after a fixed number of items and never after a run of errors: candidates in the published data lost 25 to 30 levels without the game stopping. There is no three-strikes rule.
No going back, and no penalty
There is no item navigation, no review screen and no flagging. The only way back is to fail a level and be moved down. There is no negative marking either, and structurally there cannot be: a wrong answer costs you a position and the time you spent, not points. You can pause at any time to re-read the instructions.
Playing more levels is not the strategy, and HireVue's own data says so
Almost every HireVue preparation page tells you to maximise throughput. Phrases like "accuracy per minute is the consistent theme across the cognitive games" are everywhere. HireVue's own published numbers say the opposite.
In the scoring sample reported in HireVue's peer-reviewed paper, total levels played correlated negatively with measured cognitive ability on Shapedance: r = -0.19 against the questionnaire measure, and r = -0.63 against the game-based score itself. On Numerosity it was r = 0.08, which is essentially nothing.
The reason is mechanical rather than mysterious. Losing a level is fast, because you commit to a wrong answer quickly and move on, and it knocks you down a position. So a high level count is the signature of failing a lot, not of working hard. What the model actually weights is your highest level reached and your win ratio. The highest level on Numerosity was the single most important feature in HireVue's published model.
Practically: slow down by a beat and win the level in front of you. For calibration, the real sample averaged about 28 levels in three minutes on Numerosity with 79% won, and about 24 levels in 3.3 minutes on Shapedance with 70% won. Those are HireVue's difficulty curves rather than ours, so treat them as a target pace rather than a target score.
The eleven games, and what each one actually does
Four of these are described by name in HireVue's own peer-reviewed paper, written by its own scientists, and their definitions are used here. The rest come from three independent preparation sources that agree with each other. The games originate from MindX, a London company HireVue acquired in May 2018.
Numerosity (3 min)
HireVue’s own words: a mental arithmetic task where players are shown a set of numbers and asked to identify a combination that, when summed, subtracted, multiplied or divided, yields a target result. The fastest game in the set, around nine levels a minute, and the one whose ceiling matters most to the score.
Shapedance (3.3 min)
HireVue’s own words: a mental rotation task requiring players to identify matching patterns of increasing complexity in rotated, and rotating, stimuli. Note the duration: HireVue’s paper says 3.3 minutes and every prep site says 3. The animation is the difficulty mechanism, not decoration.
Singularity (3 min)
HireVue’s own words: an attention task requiring players to identify a unique shape among increasing numbers of distractor shapes, and with a decreasing number of unique features of the target shape. Difficulty rises on two axes at once, which no prep source mentions. It is perceptual search, not rule-finding.
Pathfinder (5 min)
HireVue’s own words: a puzzle task where players move pieces in order to create a path between two endpoints. The longest of the cognitive games, with fewer and heavier levels. No source states whether a level carries its own move budget or sub-timer, so ours uses a generous move budget and says so.
Digitspan (3 min)
A sequence of numbers and letters shown briefly, then typed back, lengthening as you climb. One preparation source, and only one, reports that later levels ask for the sequence in a modified order. A reordered span is a real and well-established escalation, so we build it from six items onward and flag that it is reported rather than confirmed.
Flashback (3 min)
An n-back task: shapes appear one at a time and you say whether the current one matches the one shown n steps earlier, with n growing. HireVue does not document how a stream of yes-or-no answers maps onto levels won and lost, so ours defines a level as a block of trials at one n, won when accuracy clears a threshold. That is our design decision.
Puzzle Picture (7 min)
A target image is shown to memorise, then broken into scrambled pieces to rebuild. The longest game in the catalogue. Dual construct: visual memory for the target plus spatial problem solving to reassemble it.
E-Motions (4 min)
Faces, and a set of emotion labels to choose from. This one has right answers, which surprises people: HireVue reports that its emotional-intelligence games correlate r = .36 to .45 with GERT-S and STEM, which are performance tests with correct answers. So it is scored like an ability test, not reported like a personality profile.
Portrait (5 min, about 50 items)
Forced choice between two images with trait descriptions: pick the one that better represents you. Non-adaptive, no right answer. Our version uses illustrated cards with captions rather than photographs, because we have no image assets and would rather say so than pretend.
PortraitXT (5 min, about 50 items)
Agreement with statements, Strongly Disagree to Strongly Agree. Non-adaptive, no right answer. About six seconds an item, which is generous but real. HireVue reports these personality games correlate r = .52 to .72 with IPIP, the public Big Five item pool.
Teamchat (5 min)
A chat interface. Messages arrive as in a messaging app and you pick one of three replies, after which the conversation moves on with your choice in it. Branching, so earlier replies change later messages. Measures preferred work styles and teamwork, with no right answer.
One preparation site publishes a table of "all 13 games" adding two more, Disconumbers and Pulse. No other source mentions either name, and the same page makes other claims that HireVue's own documentation contradicts, so we have not built them. Adding a test the vendor may not sell, in order to look more complete, is how a practice suite ends up lying to a paying candidate.
The two products that are not games
Virtual Job Tryout
The real one is built for a specific job, and ours is one representative version. HireVue publishes more than forty role-specific Virtual Job Tryouts across banking, healthcare, manufacturing, sales and retail, plus general versions for individual contributors and graduates. Yours will use your job's tasks. What carries across is the structure, which is what we simulate.
Six sections in the best-documented example: situational judgement asking for both the most and least effective response, a multi-source information task, emotion recognition from recorded messages, questions about your background, a delayed recall section about those messages, and a work-style section. It also includes a realistic job preview so candidates can opt out if the job is not for them.
The delayed recall is the trap. You are given customer details in one section and asked about them two sections later, with no warning it is coming. Almost nobody does well on it first time, which is exactly why practising it once is worth doing.
Two things we cannot settle. Whether sections are individually timed: two preparation sources contradict each other and there is no first-party statement, so ours guides rather than cuts you off. And the audio: the real thing plays recordings, ours prints the transcript, because we have no recorded assets and a silent "listen to this" screen would misdescribe what you are practising.
CodeVue, or "coding assessment"
Answer to both names: the candidate help centre still has a CodeVue FAQ while the current marketing page is headed "coding assessments", so an invitation could say either.
What HireVue documents: the employer sets the language, and you are either shown a dropdown of allowed languages or told which one to use. There is an optional practice mode straight after the Begin page, which is the only product in the family with one. Retakes are the employer's call. Submissions are automatically scored, and HireVue advertises plagiarism detection, so pasting a solution is a real risk. The language list on HireVue's current page is Python, Java, Ruby, JavaScript, PHP, C++ and Perl among others.
What nobody publishes: how many problems a challenge contains, or whether it has a clock. The FAQ discusses due dates in days and never a challenge timer. Ours runs three problems with no countdown rather than inventing a deadline.
The real CodeVue does not run on mobile devices at all, and HireVue says so in its own FAQ. Our version does, so you can read the problems and think about them on a phone, but sit the real one at a desk. Ours also runs JavaScript only, because that is what a browser can execute without shipping a runtime. What transfers is the workflow, not the syntax.
What the employer sees, and what you see
You are ranked locally
HireVue's white paper states that local norming groups are used so candidates are only compared to others in their country or region. This is genuinely unusual among the vendors we cover and it matters: you are not competing against a global pool. The employer sees competency scores and candidates stack-ranked against each other.
The cut score is the employer's
HireVue publishes no pass mark. Its own research illustrates fairness analysis with a pass rate at the 33rd percentile, calling that a practical illustration of common passing rates used in practice. Treat that as one common configuration rather than a standard, and treat any site quoting a specific HireVue pass mark with suspicion.
You may get a report later
Some candidates receive a Candidate Insight Report by emailed link, covering one to three areas depending on what they sat: Working with People, Working Style and Personality, and Working with Information. HireVue states plainly that receiving it does not mean a decision has been made, and the link never expires.
Two honest notes about our own scoring. HireVue's cognitive score is a battery-level composite across whichever games you played, and it publishes no per-game percentile at all, so the per-game percentiles here are ours rather than HireVue's. And our difficulty curves are not HireVue's, so calibrating our levels against their published means is an approximation. Both are stated in the feedback you get after each game as well as here.
What HireVue publishes about whether this works
HireVue is unusual among assessment vendors in that its own scientists have published a peer-reviewed validation paper naming four of the games and listing the gameplay features used for scoring. These are its figures, not ours.
- Cognitive games against a conventional cognitive ability questionnaire: r = .50 for the two-game model, and r = .51 to .67 depending on the combination of games.
- Emotional intelligence games against GERT-S and STEM: r = .36 to .45.
- Personality games against the IPIP Big Five item pool: r = .52 to .72.
- Test-retest reliability: r = .68 over four months, N = 102.
- Adverse impact: no adverse impact on any evaluated group on age, gender or ethnicity at a 33rd-percentile cut, N = 3,107.
- Completion rate: 95% of applicants who start.
One number where HireVue disagrees with itself: candidate net promoter score is given as above 70 in the marketing white paper and as 58 in the peer-reviewed study. We quote 58, because that is the one attached to a described sample.
Practical things worth knowing before you sit it
Use Firefox or Safari, not Chrome
HireVue’s own troubleshooting article recommends Firefox or Safari on desktop, or its mobile app, because Chrome and Edge on Windows can cause distorted characters, blurry game elements and poor rendering. On a shape-discrimination task that is not cosmetic.
A dropped connection does not lose your progress
HireVue states that if your internet connection is interrupted, your game progress is saved and you can continue once it is stable. The assessment can also be taken in one go or with breaks between the games.
Take a moment between games
There is no enforced break, but HireVue’s own advice to candidates is to take a second to relax after each game so you can begin the next one fresh. On a set of three-minute clocks, arriving at the next game still thinking about the last one is expensive.
The opening levels are meant to be easy
HireVue tells candidates the game starts with some easy levels so you can get used to the format. If you are struggling in the first minute, it is almost always the instructions rather than the task, and you can pause to re-read them.
There is no practice round for the games
Unlike CodeVue, the games have no practice mode. What they do have is an instruction video before each game, replayable from the pause screen. Watching it properly is the only preparation the real product gives you.
Colours are chosen to be colour-blind friendly
HireVue states in its own paper that game colours come from colour-blind friendly palettes. Ours use a colour-universal palette for the same reason, and every game that could depend on colour also carries shape or pattern information.
Frequently Asked Questions
Is a HireVue assessment the same as a HireVue interview?
No, and this is the most common mistake candidates make with this vendor. HireVue sells game-based assessments, the Virtual Job Tryout, CodeVue coding challenges and the OnDemand video interview, all through one platform and often in one invitation. The games are three-minute puzzle tasks on a difficulty ladder. The interview records you answering questions to camera. They are marked completely differently. If your invitation names games or a Virtual Job Tryout, this page is the right one. If it only names a recorded interview, our interview practice is at /interview.
How many HireVue games will I have to play?
Fewer than eleven. The employer chooses a subset. HireVue documents a graduate assessment made of six interview questions plus three short game challenges, and its own research describes a technical graduate package of exactly three games: Numerosity, Singularity and Shapedance. Prep sources report four to six. There is no fixed battery, so practise the mechanic rather than trying to memorise a specific line-up.
How are the HireVue games scored?
On how you played, not on which answers you gave. HireVue publishes the feature list its model uses: highest level completed, levels won, levels lost, the ratio between them, and total levels played. Those features go into a regression trained to predict scores on a conventional cognitive ability test. The highest level reached on Numerosity was the single most important feature in its published model.
Should I play as many levels as possible?
No, and this is the counter-intuitive part. In HireVue’s own scoring sample, total levels played correlated negatively with measured cognitive ability on Shapedance, at r = -0.19 against the questionnaire measure and r = -0.63 against the game-based score itself. The reason is mechanical: losing a level is quick and knocks you down a position, so churning through many levels is the signature of failing a lot of them. Your ceiling and your win ratio are what the model reads.
Can you pause a HireVue game?
Yes, at any time. HireVue’s candidate help centre says so for both the desktop and the mobile versions, and pausing lets you review the instructions again and toggle the sound. Some prep sites claim there is no pause function; first-party documentation says otherwise. The clock is still the binding constraint on how far you get, because it is per game and your level resumes where it was.
Does HireVue analyse my face?
Not any more, and not in any current product. HireVue removed facial analysis from its video screening, announcing it in January 2021 and stating the feature had actually been discontinued in March 2020 after internal research found visual analysis no longer added value. The games never analysed faces at all. E-Motions shows you faces and asks you to identify the emotion; it does not look at yours.
What is a good score on a HireVue assessment?
HireVue does not publish one, because the employer sets it. What can be said honestly is that HireVue’s own peer-reviewed research illustrates its fairness analysis with a pass rate set at the 33rd percentile, describing that as a practical illustration of common passing rates used in practice. That is a common configuration, not a HireVue standard. HireVue also norms locally, so you are compared against candidates in your own country or region rather than a global pool.
Do candidates see their HireVue results?
Not during or immediately after. Some candidates later receive a Candidate Insight Report by emailed link, which gives a description of each competency area and a development tip. HireVue states explicitly that receiving the report does not indicate a decision has been made either way, and that the link never expires.
How long is a CodeVue coding challenge?
Nobody publishes this, including HireVue. Its CodeVue FAQ discusses due dates measured in days and never mentions a challenge clock, and no source we could find states a problem count either. What is documented: the employer chooses the language, there is an optional practice mode before the real challenge, retakes are the employer’s decision, and CodeVue is not supported on mobile devices at all.
What is the Virtual Job Tryout?
A job simulation rather than a test. HireVue publishes more than forty role-specific versions across banking, healthcare, manufacturing, sales and retail, so yours is built around the job you applied for. The best-documented example runs six sections: situational judgement, a multi-source information task, emotion recognition from audio, questions about your background, a delayed recall section about the audio you heard two sections earlier, and a work-style section. It came to HireVue through Modern Hire, which got it from Shaker International, which built it in 2002.
Companies That Use HireVue
HireVue is a US-headquartered vendor whose assessments sit at the first gate of high-volume and graduate hiring, most often in banking, healthcare, retail, manufacturing and airlines. The games, the Virtual Job Tryout and the recorded video interview arrive from the same platform and often in the same invitation, so confirm which of the three you have been sent.
Philips
Healthcare TechnologyGame-based assessment alongside the video interview in global early-careers hiring.
View Philips guide →Delta Air Lines
AirlinesVirtual Job Tryout, reported for flight attendant and airport customer service hiring.
View Delta Air Lines guide →HireVue is far better known for its recorded video interview than for its assessments, and many employers use only that. An employer appearing here has evidence for the assessment product specifically. Practice for a recorded interview lives at /interview.
Practice by Test Type
Compare the same format across every provider that uses it, and see what each one is really measuring.
Numerical Reasoning Test Practice
Tables, charts, percentages and ratios against the clock. The most common screening test on graduate schemes and analyst roles.
Logical Reasoning Test Practice
Shapes, sequences and flowcharts instead of words or numbers. Inductive, deductive, abstract, diagrammatic and spatial in one place.
Situational Judgement Test Practice
Realistic workplace scenarios scored against an employer competency framework. No maths, no time pressure, and a scoring key you can reason about.
Personality Test Practice for Job Applications
No right answers, which is exactly why they unsettle people. Learn the traits, the consistency checks and the formats before you sit one for real.
Ready to Start Practising?
Sign up for free to try 2 tests, or unlock all 13 HireVue simulations with unlimited practice and full feedback.