How to Train for Hour Memory Events: Numbers and Cards

SYSTEM → PACE → CAPACITY → ENCODE → REVIEW → RECALL → RECOVER

Hour Numbers and Hour Cards use the same encoding techniques as their five-minute cousins, but they are not five-minute events repeated twelve times. They are endurance events: success depends on a sustainable encoding pace, enough memory-palace capacity, deliberate review, controlled recall, and a plan for handling errors and fatigue — not just on how fast you can form one image.

This guide covers all of that: what these events actually involve, why an hour breaks strategies that work fine for five minutes, how to build the capacity and review habits a long session needs, and how to train toward the full format using this site's actual Hour Numbers and Hour Cards drills.

Quick answer

Effective hour-event training comes down to ten things:

1. A reliable, well-drilled encoding system for digits or cards.
2. A large enough supply of prepared memory-palace journeys.
3. A sustainable encoding pace, not your fastest five-minute pace.
4. Planned review points built into the memorization period.
5. Session duration increased progressively, not jumped to 60 minutes.
6. Recall trained separately from memorization.
7. A way to manage errors without losing time on any single one.
8. Full-length sessions rehearsed periodically, not every week.
9. Performance tracked by time segment, not just as a final score.
10. Physical and mental preparation for a session that can run several hours.

What are Hour Numbers and Hour Cards?

Hour Numbers and Hour Cards are the long-duration counterparts to Speed Numbers and Speed Cards. All four events use the same basic skill — converting digits or cards into images and placing them on a route — but Hour Numbers and Hour Cards extend the memorization period from a few minutes to a full hour, and extend recall to match.

"Hour event" describes the memorization period, not necessarily the whole event. At World level, the World Memory Sports Council's own description of the sport notes that disciplines such as these can run a one-hour memorization period with a two-hour recall period — three hours in total, not one. Lower competition levels use shorter periods for both phases.

Memorization and recall are also different skills applied to the same material: memorization is encoding new information as fast as you reliably can, while recall is retrieving and accurately writing down everything you encoded, often much later and under its own time pressure. An hour-long event stresses both, separately, which is a large part of why it behaves so differently from a short sprint.

Why an hour is different from five minutes

In a five-minute event you can often encode continuously, use one short journey, tolerate very high intensity, delay review entirely, and rely on the material still being fresh a few minutes later. None of that holds for an hour.

None of this means every athlete needs an identical strategy — pace, review ratio, and route design are all personal. But every hour-event plan has to account for these factors somehow; the short event mostly lets you ignore them.

The three-hour problem

At World level the arithmetic is blunt: 60 minutes of memorization plus 120 minutes of recall is 180 minutes. The athlete is managing a one-hour encoding workload and a two-hour recall workload back to back, with the same concentration, the same physical comfort, and the same route discipline required for both.

The goal isn't merely to survive the memorization hour. You also have to preserve enough cognitive and physical capacity — clear images, an intact route structure, legible handwriting or accurate typing, and the concentration to keep going — for two more hours of recall afterward. A plan that maximizes what you memorize while leaving nothing for recall will still score badly.

Know your target pace

Records give a sense of scale, not a beginner target. The Hour Numbers world record is 4,620 digits (Ryu Song I, 2019 World Memory Championships, Wuhan — per Guinness World Records), which works out to exactly 77 digits a minute for the full hour. The Hour Cards record is 2,530 cards (Kim Surim, same championship — also Guinness) — about 42.2 cards a minute, or roughly a pack every 74 seconds.

Do not use these as an early goal. The number that actually matters is your own, and the only way to find it is to measure it directly:

YOUR TARGET PACE = THE FASTEST PACE YOU CAN SUSTAIN AT ACCEPTABLE ACCURACY, WITH YOUR REVIEW PLAN INTACT

Useful units to track as you go: digits per minute, cards per minute, seconds per image, images per minute, packs per hour, and rows per minute. All of them describe the same underlying thing — how fast you reliably convert material into placed images — from a different angle.

Find your sustainable pace

Run a short series of timed sessions at increasing duration, and track how your pace and accuracy hold up as the duration grows:

Run these in Hour Numbers or Hour Cards Custom mode, scaling the digit count or pack count and both timers down to match. The goal isn't to maximize the five-minute row — it's to see whether your five-minute pace, and your accuracy at that pace, actually survives to 60 minutes.

Memory systems for Hour Numbers

Any digit system works for Hour Numbers in principle; what changes at hour-event scale is how much route capacity each one costs.

• Two-digit Major System. Each pair of digits becomes one image. With two images placed at each memory-palace location, a locus holds 4 digits — so 1,000 digits needs 250 locations. See the Major System guide.
• Three-digit systems. Higher compression per image, at the cost of a much larger trained image vocabulary (up to 1,000 entries instead of 100) that has to be equally automatic before it helps.
• PAO for numbers. A person-action-object system can encode multiple digits per scene, depending on how it's built; see the PAO System guide for the trade-offs.

None of these is universally fastest — the right choice depends on which system you already have well drilled and how much time you've invested in the image vocabulary it requires.

Memory systems for Hour Cards

• One image per card. Conceptually simple, but a full 52-card pack needs 52 images and 52 loci — the most route-hungry option at hour-event scale.
• Two-card images. Pairing cards halves the count: 52 cards become 26 images per pack.
• PAO. With a person-action-object system, three cards combine into one scene: 52 cards split into 17 complete three-card groups plus one remaining card, for roughly 18 scenes per pack — the same figure this site's own deck-memorization guide and PAO guide use.

Card-system automation matters even more in a one-hour event than in a five-minute one, simply because the same lookup happens dozens of times over instead of a handful — any hesitation in retrieving a card's image compounds across 59 packs.

Build enough memory-palace capacity

Route capacity is a hidden bottleneck: running out of locations mid-session is an entirely avoidable failure, but only if it's planned for in advance. The numbers below follow directly from the encoding architecture chosen above — they are examples under that architecture, not universal requirements.

Numbers, two-digit images at two per locus (4 digits/locus): 1,000 digits needs 250 loci; the 4,620-digit world record needs about 1,155 loci; this site's 5,560-digit Standard sheet needs about 1,390.

Cards: with one image per two cards, a pack needs 26 loci — 59 packs need about 1,534. With an 18-scene PAO pack, that drops to about 1,062 loci for the same 59 packs.

Whichever architecture you use, the practical rule is the same: work out the minimum your material requires, then build meaningfully more than that minimum before you need it.

How many memory palaces do you need?

A few terms are worth keeping distinct: a journey or route is a fixed path through a place you know; a locus is one stop on that route; and a long event is naturally divided into blocks — pages, for numbers, or packs, for cards.

A common approach is to assign a different journey to each block, so no single route has to hold the whole event:

1. 1Page 1 → Journey A
2. 2Page 2 → Journey B
3. 3Page 3 → Journey C
4. 4Page 4 → Journey D
5. 5Page 5 → Journey E

The same idea applies to cards, with one journey per pack or per small group of packs. This isn't a required structure — the application and competition format don't mandate it — but it keeps a weak or exhausted route from contaminating material you haven't reached yet, and it avoids reusing a route so soon that yesterday's images are still sitting on it, creating interference with today's.

Route design for long events

A route that holds up for an hour of new material tends to be:

• Familiar, so you're not spending attention navigating it.
• Strictly ordered, with no ambiguity about what comes next.
• Visually distinct at every locus, so loci don't blur together.
• Easy to traverse, with enough loci for the block it's assigned.

A route that tends to fail under load has the opposite problems:

• Repeated or visually similar locations.
• Ambiguous ordering, or a path that crosses itself.
• Locations placed too close together to distinguish under pressure.
• Awkward transitions between loci that break your rhythm.

Build and rehearse routes with Palace Builder until walking them is automatic; see the method of loci for the underlying technique.

Don't memorize for 60 minutes without a review plan

Material memorized at minute 5 has sat in memory for roughly 55 minutes by the time encoding ends. Material memorized at minute 58 is still fresh. That gap in age is the whole reason review scheduling exists — and the reason a five-minute strategy of "encode, then recall immediately" doesn't transfer to an hour.

• Page-based review. After finishing a page or pack, review older material before moving on.
• Block-based review. Memorize a whole block, then deliberately review it before starting the next.
• Rolling review. Keep encoding new material while periodically revisiting older blocks.
• Minimal review. For a strong system and well-drilled routes, reduce review and maximize new material instead.

The right balance among these is individual, not fixed — see the experiment below for how to find yours.

Review vs. new material

New material increases your potential score; review protects material you've already earned. Too much new material and early images fade before recall; too much review and you simply memorize less. There is no universal ratio here — a claim like "spend 80% encoding, 20% reviewing" isn't supported by anything verifiable, so this guide won't assert one. The right balance is something you find, not something you're told.

How to discover your review schedule

Run two sessions of the same duration and compare outcomes, not just effort:

• Session A — minimal review. Encode continuously; review only at the very end.
• Session B — regular review. Build in review checkpoints throughout.

Compare amount memorized, amount actually recalled, final score, and where errors landed by time segment. The objective is to maximize your final recalled score, not the raw amount of material you touched — a session that memorizes more but recalls a smaller fraction of it correctly is the worse result.

Pacing the memorization hour

Divide the hour into checkpoints and evaluate at each one:

ENCODE → PLACE → REVIEW → CONTINUE → RECALL → ANALYZE → ADJUST

At each of the 0–10, 10–20, 20–30, 30–40, 40–50, and 50–60 minute marks, check your pace, your accuracy, your fatigue level, how much route you've used, and how much review you owe yourself. This guide deliberately doesn't prescribe exact targets for any of these — the point of checking is to find your own sustainable pace experimentally, using the test earlier in this guide.

Avoid the five-minute pace trap

A five-minute sprint rewards maximum intensity, and that habit is hard to switch off. Carrying five-minute intensity into an hour-long session tends to produce image-quality deterioration, route errors, attention lapses, reduced review, accumulating fatigue, and — often only visible at recall — a collapse in the last third of the material. The objective for an hour event is maximum sustainable performance, not maximum short-term speed.

Error management during memorization

When an image feels weak or a sequence feels uncertain mid-session:

1. Identify the uncertainty quickly — don't let it linger unnoticed.
2. Repair it immediately if the repair is cheap.
3. If the repair is expensive, note it mentally and keep moving.
4. Never let one bad item consume the time needed for many good ones.
5. Return to it later only if your review schedule allows.

The distinction that matters is between local uncertainty — one weak image in an otherwise solid route — and global collapse, where panic over one item starts costing you several more. Protect the rest of the hour over any single item.

Error management during recall

This site's Hour Numbers and Hour Cards drills both use the same row-based World Memory Sports Council scoring, verified directly against the scoring implementation:

The practical implication follows directly from "2 or more mistakes scores zero, but nothing is attempted scores zero anyway": if a row or pack is genuinely lost, continuing to spend recall time on it can be worse than leaving it blank and moving on. A blank row costs nothing; a wrong guess costs nothing extra either, but only if it's your one mistake in that row — a second guess that's also wrong drops the whole row to zero.

This is sharpest in Hour Cards: two swapped cards count as two separate errors, so a guessed swap has an even chance of scoring a full pack or zero. Leaving one of the two positions blank instead is a single mistake — a guaranteed half-pack score rather than a coin flip. Whether that trade is worth it depends on how confident you actually are; this guide isn't telling you to always leave positions blank, only that the scoring rules make the trade-off calculable.

Recall pacing

Recall isn't just writing down what you remember — it can be its own bottleneck. For numbers, that means maintaining row alignment, tracking page boundaries, and avoiding transcription slips over a long writing session. For cards, it means preserving pack order, avoiding transpositions, and moving cleanly between packs without losing your place.

At the world-record pace, two hours of recall for the record numbers works out to about 39 digits a minute across the whole period; for cards, about 21 a minute. Both are far slower than memorization pace — recall has its own rhythm, and it's worth checking the clock at planned points so the final page or pack isn't rushed.

Train recall as a separate skill

Strong encoding and strong recall output are genuinely different skills, and it's entirely possible to have one without the other. Worthwhile recall-specific drills include:

• Short memorization followed by unusually long recall, to practice unhurried, accurate output.
• Recall under tighter time pressure than usual.
• Partial recall — memorize a full session but only recall part of it, picking up a route mid-way.
• Row-by-row or pack-by-pack recall drills, isolating output speed from encoding speed.
• Deliberate transcription-accuracy checks, independent of whether the memory itself was correct.

The last 10 minutes of memorization

There isn't one universal strategy for the closing minutes — but there are trade-offs worth deciding on in advance rather than mid-session: reviewing your oldest, most fade-prone material; declining ambitious new material if it would destabilize routes you've already secured; protecting blocks you're already confident in rather than chasing a few more; and mentally shifting toward the recall mindset before the clock actually changes phase.

Training progression

A recommended training progression — not an official competition rule:

1. 1
   Build the systemA digit or card system you don't have to think about · The Major System, PAO, or a comparable encoding

   How to memorize numbersHow to memorize a deck of cards
2. 2
   Find sustainable paceTest 5, 10, 20, 30, 45, 60 minutes · Find where accuracy starts to fall
3. 3
   Build capacity and routesMore memory-palace locations than the minimum · Familiar, ordered, visually distinct routes

   Method of lociPalace Builder
4. 4
   Add review and enduranceA planned review schedule · Progressively longer sessions
5. 5
   Simulate and trainFull 60-minute memorization plus extended recall · Standard, National, and Custom modes

   Train Hour NumbersTrain Hour Cards

In terms of raw duration, that's roughly: 5 minutes to learn the system, 10 to develop pace, 20 to introduce review scheduling, 30 to train endurance, 45 to stress-test route capacity, 60 for a full memorization session, and finally a full competition simulation — 60 minutes of memorization plus extended recall.

Custom mode training

Both drills' Custom mode lets you change digit count or pack count, memorization time, and recall time independently. The tunable ranges are 40–6,000 digits (in steps of 40) for Hour Numbers and 1–60 packs for Hour Cards, with both timers adjustable from 1 minute up to 60 minutes memorizing and up to 2 hours recalling. A fresh Custom run starts from a small default — 150 digits or 2 packs, 10 minutes each way — before you move any sliders. Custom runs can also be paused, unlike Standard mode, which keeps results comparable by disallowing it.

Use Custom mode to isolate a specific weakness rather than always running the full format:

• Weak endurance: increase duration while holding pace roughly constant.
• Weak recall: increase recall time relative to memorization time.
• Weak encoding: reduce the amount of material and focus purely on image quality.
• Weak review: deliberately add planned review intervals and measure the effect.
• Weak route capacity: increase the material until you feel the journey count run short.

Standard mode

Standard mode in Hour Numbers runs exactly the World Memory Championship timing: 5,560 digits, in rows of 40 across 6 pages of 25 rows (139 rows total), with 60 minutes to memorize and 120 minutes to recall, in a single attempt. The digit count isn't an official competition figure — it's this site's own sizing rule, the world record plus 20%, rounded up to whole rows, so nobody ever reaches the end of the sheet.

Standard mode in Hour Cards gives a pool of 59 separately shuffled packs of 52 cards, on the same record-plus-20%-in-whole-packs basis, with the same 60-minute memorization and 120-minute recall, also a single attempt. Hour Cards additionally offers National and International presets that memorize from the same 59-pack pool but shorten the clock to 10/30 and 30/60 minutes respectively — the tiered timing the published WMSC rulebook documents for the Random Cards discipline. Hour Numbers has no equivalent National or International preset button on this site; the WMSC's published rulebook doesn't include a Numbers-specific section, so only the World-level 60/120 timing is used here.

Training metrics

Worth tracking across a session, for numbers:

Digits memorized · digits recalled · final score · digits per minute · perfect rows · one-error (half-score) rows · zero-score rows · attempted rows · decision points (the WMSC tie-break: correctly placed digits inside zero-scoring rows) · performance by 10-minute block · recall speed.

And for cards:

Cards memorized · packs completed · final score · cards per minute · perfect packs · one-error packs · zero-score packs · attempted packs · decision points · partial final pack · recall speed.

For endurance specifically, track performance at 10, 20, 30, 40, 50, and 60 minutes — the goal is to find where performance actually begins to deteriorate, which is rarely obvious from a single final score alone.

Analyze performance by time segment

Some patterns and what they tend to suggest (not a diagnosis, just a starting point for adjustment):

• Pace stays stable but recall drops → likely a review problem.
• Recall stays strong but pace collapses → likely an encoding-endurance problem.
• Both collapse after the same point every session → likely pacing, fatigue, or concentration — worth investigating as a pattern, not treating as a one-off bad day.

Physical and environmental preparation

Keep this practical: comfortable posture and desk height, stable and consistent lighting, hydration, minimizing avoidable distractions and interruptions, a comfortable writing or typing setup, sensible breaks during training (not during a scored attempt), and — where possible — practicing at the time of day a real competition would run.

Concentration management

Sustained attention over an hour is trained like anything else. Practical habits: a fixed pre-session routine, notifications off, a consistent training environment, a review schedule decided before you start rather than improvised mid-session, short mental resets rather than long ones, and simply returning attention to the next chunk after an error rather than dwelling on it.

Competition-day preparation

Before the event: confirm your system is fully automatic, confirm your routes are prepared and rehearsed, know your target pace, have practiced full-length recall, be appropriately rested, have your writing materials ready if applicable, and know the scoring and page/pack structure in advance.

During memorization: hold your sustainable pace, track how much route you've used, review according to your plan, protect material you're already confident in, and don't let one error trigger a cascade of panic.

During recall: follow the order you memorized in, respect page/row/pack boundaries, skip material that's genuinely lost when the scoring makes that the rational choice, preserve time for later material, and avoid guessing where a blank scores better.

How to simulate a real competition

1. Warm-up. A short, low-stakes run to settle into rhythm.
2. 60-minute memorization. The full Standard-mode session.
3. Short transition. A brief break, mirroring the gap before recall actually opens.
4. 120-minute recall. The full recall window, run without stopping.
5. Performance analysis. Review the time-segment breakdown above and adjust your plan.

Occasionally running the full duration matters — it's the only way to test whether pace, capacity, and review actually hold together for three hours — but not every session needs to be a full simulation.

A sample weekly training plan

An example schedule, not an official training requirement:

• Monday — Numbers: encoding speed drills.
• Tuesday — Cards: image automation drills.
• Wednesday — Short endurance and review experiments.
• Thursday — Numbers: a 30–45 minute session.
• Friday — Cards: a 30–45 minute session.
• Saturday — A full simulation, alternating which event.
• Sunday — Review errors from the week; light training.

Four-week progression

• Week 1 — System and pacing. Focus entirely on reliable, automatic encoding.
• Week 2 — Endurance. 10–30 minute sessions, extending gradually.
• Week 3 — Review and long sessions. 30–45 minute sessions with a deliberate review schedule.
• Week 4 — Competition simulation. A full 60-minute memorization plus extended recall, with recovery days built in around it.

Hour Numbers vs. Hour Cards

Neither is inherently harder — difficulty depends heavily on which system and route habits you already have.

Training both events together

What transfers between Hour Numbers and Hour Cards: route design, review scheduling, pacing discipline, concentration management, error-handling strategy, and recall discipline. What doesn't transfer directly: the number encoding vocabulary, the card encoding vocabulary, and the exact recall mechanics of each. Most athletes training both keep separate encoding systems while sharing the same higher-level endurance strategy across them.

Common training mistakes

1. Training only five-minute sprints.
2. Never training recall on its own.
3. Memorizing for long sessions with no review plan.
4. Running out of memory-palace locations mid-session.
5. Reusing a journey too soon, before yesterday's images have faded.
6. Changing mnemonic systems constantly instead of automating one.
7. Chasing speed before the conversion is genuinely automatic.
8. Ignoring transcription errors that have nothing to do with memory.
9. Never measuring performance by time segment.
10. Practicing only when motivated, rather than on a schedule.
11. Running full three-hour simulations too frequently.
12. Ignoring physical comfort during long sessions.
13. Treating one bad session as proof the whole system is broken.
14. Failing to analyze exactly where performance deteriorated.

Troubleshooting

The most important principles

1. Build the system before chasing speed.
2. Train sustainable pace, not five-minute pace.
3. Build more route capacity than you think you need.
4. Review strategically, not by habit or not at all.
5. Treat recall as a separate skill from memorization.
6. Protect time from low-value errors.
7. Analyze exactly where performance deteriorates.
8. Increase duration gradually.
9. Simulate the full event occasionally, not constantly.
10. Optimize final recalled score, not raw memorization volume.

An athlete's ceiling in these events is ultimately set by four things together — encoding speed, memory capacity, review capacity, and recall capacity — not any one of them alone.

FAQ

Train it

Reading about pacing and review is no substitute for testing them under a clock. Run a Custom session at a fraction of the full format, find your sustainable pace, and build up from there:

Build your endurance with a timed session, and once the basics hold, work up through Speed Numbers and Speed Cards at the other end of the same skill.

Related techniques

• How to Memorize Numbers — building the digit system this guide assumes.
• How to Memorize a Deck of Cards — the card-system equivalent.
• Method of Loci — building and using the routes this guide relies on.
• PAO System — the person-action-object build used for higher-compression encoding.
• Major System — the two-digit encoding used in the capacity calculations above.
• Palace Builder — for designing and rehearsing the routes a long event needs.
• Spaced Repetition — the same review-scheduling principle, compressed into an hour.