TOP games in terms of "time in the game"
1) What exactly we measure
Base metrics:
Session Definition:
2) Data cleaning and quality
We exclude AFK tails: pauses> 10 minutes inside the session are not taken into account in the duration.
Auto-spin filter: If there are> 500 non-click spins in a row, limit the contribution to "in-game time" (e.g. consider 50%).
Remove anomalies: sessions> 240 minutes check; bots - by the absence of scrolling/clicks/transitions.
We break down the report by device (mobile/desktop), geo, traffic source, rate (micro/standard/high roll).
Recommended period: last 30 days + 7-day sidebar for "pulse" trends.
3) Ranking model (transparent and reproducible)
TIME\_ SCORE summary score (0-100) - linear composition of normalized indicators within the cohort (type: Min-Max or Z-score; use the same type for all factors):
Penalty for high bounce: if the session share is <60 sec> 25%, subtract up to 5 points proportionally.
Why so:
4) Why specific slots gain "in-game time"
Retention design patterns:
5) TOP games by "in-game time" (approximate editorial matrix)
1. Gates of Olympus (Pragmatic Play)
Patterns: cascades + global multiplier, frequent micro-highlights.
For whom: mass audience, average bankroll.
Plus: "there is always a chance of an x-cartoon," the pace does not tire.
2. Sweet Bonanza (Pragmatic Play)
Patterns: clusters + multipliers in freespins; simple readability.
For whom: beginners/mid-core.
Plus: a stable cycle of "events," high SPU.
3. Big Bass Bonanza (series, Pragmatic Play)
Patterns: collection collection, multi-stage levels of freespins.
For whom: players who love progress and "finish."
Plus: strong RR D7 thanks to the series (Big Bass, Christmas, Splash, etc.).
4. Book of Dead (Play’n GO)
Patterns: classics of "books," clear expectations of a bonus.
For whom: old school/book fans.
Plus: high median session due to recognition and pace.
5. Starburst (NetEnt)
Patterns: expanding Wild, frequent micro-events.
For whom: beginners and mobile.
Plus: low fatigue, high TP100 comfort.
6. Money Train 4 (Relax Gaming)
Patterns: A "collective" bonus with character synergies.
For whom: experienced, patient, high rollers.
Plus: strong P75 due to long "bonus hunting."
7. Wanted Dead or a Wild (Hacksaw Gaming)
Patterns: High-risk bonuses with "jerk" potential.
For whom: experienced, fans of the "gambling peak."
Plus: high RR in a loyal niche with competent risk communication.
8. Sugar Rush (Pragmatic Play)
Patterns: cumulative multipliers on the grid.
For whom: mid-core.
Plus: "domino effect" holds attention, high SPU.
9. Razor Shark (Push Gaming)
Patterns: hidden characters/multipliers, "tense waiting."
For whom: lovers of "tension" and volatility.
Plus: high P75 with adequate ASD.
10. The Dog House Megaways (Pragmatic Play)
Patterns: Sticky vs Raining is a choice of bonus style.
For whom: a wide audience, mobile.
Plus: frequent micro-reasons "up to one bonus."
11. Bonanza Megaways (BTG)
Patterns: cascades + line extensions.
For whom: players who value "forward" pace.
Plus: even ASD, strong median.
12. Book of Ra Deluxe (Novomatic/Greentube)
Patterns: "book" classics with nostalgia.
For whom: old school/land migrants online.
Plus: a high RR D7 in its audience.
6) How to arrange the page of the section "The most popular slots at the moment"
TOP table (minimum set of fields):
Filters on the page: Mechanics, Volatility, Device, Bankroll, Geo, Beginners/Experienced.
Blocks: "ASD Hits of the Week," "Median Leaders," "Best for Beginners (Low Volatility)," "Megaways P75 Leaders."
7) Segmentation by audience (for internal recommendations)
Beginners/micro bets: Starburst, Sweet Bonanza, fruit classics (low variance, clear rules).
Mid-core: Gates of Olympus, Sugar Rush, Dog House Megaways (event without overload).
High rollers/experienced: Money Train 4, Wanted Dead or a Wild, Razor Shark (ready for dispersion).
Lovers of "books": Book of Dead, Book of Ra Deluxe (the usual loop of waiting for a bonus).
Mobile: projects with clear TP100 and understandable animations (Starburst, Big Bass series).
8) Calculation practice (how to reproduce quickly)
Session: group by 'user _ id', 'game _ id', breaks of 30 minutes.
ASD: `avg(session_duration_minutes)`; Median/P75: percentiles by 'session _ duration _ minutes'.
SPU: `count(distinct session_id) / count(distinct user_id)` за 30д.
RR D7: 'user _ id' share with re-session in\[ 2; 7] day from the first.
TP100: 'avg (time_spent_seconds/( spins/100)) ', excluding AFK.
9) Game Card Editorial Rules
1 paragraph of essence (mechanics + "holds for what").
Mini-table: RTP, Volatility, Min/Max rate, Bonus trigger, Max multiplier.
Block "Who fits" (beginner/mid-core/high-roller).
Internal links: provider ↔ mechanics ↔ similar.
FAQ 3-4 short responses (volatility, bonus frequency, tempo).
10) Rating update and methodology integrity
Update frequency: weekly for the "Pulse of the week" widget, monthly - the main TOP.
Transparency: explain the formula and filters (AFK, autospin, outliers).
Stability: Show a minimum 30-day window + 7-day cut.
Compliance: 18 +, responsible play. No win promises.
11) Frequent Questions (FAQs)
Why is the median more important than the average? It is more resistant to rare "marathons."
Why P75? Shows the quality of long sessions without tail distortion.
Why are some hits lower in the TOP? High volatility may lower the ASD/median, but appeal to parts of the audience.
Do I need to consider the demo? Yes, but separately, so as not to dilute the performance of a real game for money.
12) The bottom line
Retention in slots is objectively measured not by a single digit, but by a package of metrics. By applying the transparent TIME\_ SCORE formula and uniform data cleaning rules, you get an honest TOP for "time in game," which can be updated on a schedule, filtered for audiences and used as the central block of the "Most popular slots so far" section.
Base metrics:
- ASD (Average Session Duration) - average session duration, min.
- Median SD - median session duration (resistant to outliers).
- P75 SD - 75th percentile of duration (quality of long sessions).
- SPU (Sessions per User) - sessions per user per period.
- RR D1/D7 is the proportion of re-entries on Day 1 and Day 7.
- Spins/Session - average number of spins per session.
- TP100 - time per 100 spins, sec (rhythm of the game, "pace").
Session Definition:
- The session is disconnected due to inactivity for ≥30 minutes or upon explicit exit (fix any of the options consistently).
2) Data cleaning and quality
We exclude AFK tails: pauses> 10 minutes inside the session are not taken into account in the duration.
Auto-spin filter: If there are> 500 non-click spins in a row, limit the contribution to "in-game time" (e.g. consider 50%).
Remove anomalies: sessions> 240 minutes check; bots - by the absence of scrolling/clicks/transitions.
We break down the report by device (mobile/desktop), geo, traffic source, rate (micro/standard/high roll).
Recommended period: last 30 days + 7-day sidebar for "pulse" trends.
3) Ranking model (transparent and reproducible)
TIME\_ SCORE summary score (0-100) - linear composition of normalized indicators within the cohort (type: Min-Max or Z-score; use the same type for all factors):
- ```
- TIME_SCORE = 0. 35·N(ASD) + 0. 25·N(Median SD) + 0. 15·N(P75 SD)
- + 0. 15·N(SPU) + 0. 10·N(RR D7) – Penalty
- ```
Penalty for high bounce: if the session share is <60 sec> 25%, subtract up to 5 points proportionally.
Why so:
- ASD - "mass" of retention.
- The median is robustness in a typical session.
- P75 is the quality of long sessions without fighting tails.
- SPU - real return during the period.
- RR D7 - weekly loyalty as a marker of "non-attendance."
4) Why specific slots gain "in-game time"
Retention design patterns:
- Average/below average volatility → frequent micro-wins, low "dead series."
- Step-by-step bonuses/progress (counters, multipliers, symbol collections).
- Cascades/clusters → series of events in one back, "continuation effect."
- Sticky mechanics (sticky Wild, collector symbols) → waiting to be "completed."
- Megaways/multipliers → hope for a "breakthrough" with a stable base game.
- Clear rhythm (TP100 in the "comfortable" range of 65-90 seconds) on mobile.
5) TOP games by "in-game time" (approximate editorial matrix)
💡Below - ready-made cards for the page. Metrics are stamped from your BI, reasons for retention - as clues for copyright. Order - by TIME\_ SCORE.
1. Gates of Olympus (Pragmatic Play)
Patterns: cascades + global multiplier, frequent micro-highlights.
For whom: mass audience, average bankroll.
Plus: "there is always a chance of an x-cartoon," the pace does not tire.
2. Sweet Bonanza (Pragmatic Play)
Patterns: clusters + multipliers in freespins; simple readability.
For whom: beginners/mid-core.
Plus: a stable cycle of "events," high SPU.
3. Big Bass Bonanza (series, Pragmatic Play)
Patterns: collection collection, multi-stage levels of freespins.
For whom: players who love progress and "finish."
Plus: strong RR D7 thanks to the series (Big Bass, Christmas, Splash, etc.).
4. Book of Dead (Play’n GO)
Patterns: classics of "books," clear expectations of a bonus.
For whom: old school/book fans.
Plus: high median session due to recognition and pace.
5. Starburst (NetEnt)
Patterns: expanding Wild, frequent micro-events.
For whom: beginners and mobile.
Plus: low fatigue, high TP100 comfort.
6. Money Train 4 (Relax Gaming)
Patterns: A "collective" bonus with character synergies.
For whom: experienced, patient, high rollers.
Plus: strong P75 due to long "bonus hunting."
7. Wanted Dead or a Wild (Hacksaw Gaming)
Patterns: High-risk bonuses with "jerk" potential.
For whom: experienced, fans of the "gambling peak."
Plus: high RR in a loyal niche with competent risk communication.
8. Sugar Rush (Pragmatic Play)
Patterns: cumulative multipliers on the grid.
For whom: mid-core.
Plus: "domino effect" holds attention, high SPU.
9. Razor Shark (Push Gaming)
Patterns: hidden characters/multipliers, "tense waiting."
For whom: lovers of "tension" and volatility.
Plus: high P75 with adequate ASD.
10. The Dog House Megaways (Pragmatic Play)
Patterns: Sticky vs Raining is a choice of bonus style.
For whom: a wide audience, mobile.
Plus: frequent micro-reasons "up to one bonus."
11. Bonanza Megaways (BTG)
Patterns: cascades + line extensions.
For whom: players who value "forward" pace.
Plus: even ASD, strong median.
12. Book of Ra Deluxe (Novomatic/Greentube)
Patterns: "book" classics with nostalgia.
For whom: old school/land migrants online.
Plus: a high RR D7 in its audience.
💡Note: take the final order and metric values from your logs.
6) How to arrange the page of the section "The most popular slots at the moment"
TOP table (minimum set of fields):
| Game | Provider | Mechanics | Volatility | ASD (min) | Median | P75 | SPU | RR D7 | Note | |
|---|---|---|---|---|---|---|---|---|---|---|
| ---------------- | -------------- | ----------------- | ------------- | --------: | -----: | --: | --: | ----: | ----------------- | |
| Gates of Olympus | Pragmatic Play | Cascades/Multipliers | Mid-high | - | - | - | - | - | - | "jerk" multiplier |
| … | … | … | … | … | … | … | … | … | … |
Filters on the page: Mechanics, Volatility, Device, Bankroll, Geo, Beginners/Experienced.
Blocks: "ASD Hits of the Week," "Median Leaders," "Best for Beginners (Low Volatility)," "Megaways P75 Leaders."
7) Segmentation by audience (for internal recommendations)
Beginners/micro bets: Starburst, Sweet Bonanza, fruit classics (low variance, clear rules).
Mid-core: Gates of Olympus, Sugar Rush, Dog House Megaways (event without overload).
High rollers/experienced: Money Train 4, Wanted Dead or a Wild, Razor Shark (ready for dispersion).
Lovers of "books": Book of Dead, Book of Ra Deluxe (the usual loop of waiting for a bonus).
Mobile: projects with clear TP100 and understandable animations (Starburst, Big Bass series).
8) Calculation practice (how to reproduce quickly)
Session: group by 'user _ id', 'game _ id', breaks of 30 minutes.
ASD: `avg(session_duration_minutes)`; Median/P75: percentiles by 'session _ duration _ minutes'.
SPU: `count(distinct session_id) / count(distinct user_id)` за 30д.
RR D7: 'user _ id' share with re-session in\[ 2; 7] day from the first.
TP100: 'avg (time_spent_seconds/( spins/100)) ', excluding AFK.
9) Game Card Editorial Rules
1 paragraph of essence (mechanics + "holds for what").
Mini-table: RTP, Volatility, Min/Max rate, Bonus trigger, Max multiplier.
Block "Who fits" (beginner/mid-core/high-roller).
Internal links: provider ↔ mechanics ↔ similar.
FAQ 3-4 short responses (volatility, bonus frequency, tempo).
10) Rating update and methodology integrity
Update frequency: weekly for the "Pulse of the week" widget, monthly - the main TOP.
Transparency: explain the formula and filters (AFK, autospin, outliers).
Stability: Show a minimum 30-day window + 7-day cut.
Compliance: 18 +, responsible play. No win promises.
11) Frequent Questions (FAQs)
Why is the median more important than the average? It is more resistant to rare "marathons."
Why P75? Shows the quality of long sessions without tail distortion.
Why are some hits lower in the TOP? High volatility may lower the ASD/median, but appeal to parts of the audience.
Do I need to consider the demo? Yes, but separately, so as not to dilute the performance of a real game for money.
12) The bottom line
Retention in slots is objectively measured not by a single digit, but by a package of metrics. By applying the transparent TIME\_ SCORE formula and uniform data cleaning rules, you get an honest TOP for "time in game," which can be updated on a schedule, filtered for audiences and used as the central block of the "Most popular slots so far" section.
