War of Genesis: Idle Loot is a free Windows program that parks a 3D hero in a transparent strip on your desktop and lets it fight, loot and level while you do something else. New Normal Soft Inc. opened Early Access on 10 September 2026, reached a peak of 52,835 concurrent players on 19 September — SteamDB’s own post on the game’s news feed carries that figure — and left Early Access with the 7 October update. Steam’s Most Played chart had it near the hundredth slot in late October — the desk’s 29 October snapshot reads rank 100 at 14,789 in game, and a poll of the current-players endpoint returned 15,045.
This site already has an idle-game primer, the TBH: Task Bar Hero loop, and it covers a different animal: a taskbar pixel RPGRPG. Role-Playing Game — a game built around character progression, story, and player choices. whose payouts and prestige ladder are documented across forty changelog posts. What that record cannot give you is the shape of a stage — which pool a drop comes from, what a synthesis attempt is worth, and which October patch already invalidated older advice.
One limit, stated up front: none of the 23 posts in the Steam news feed between 12 August and 21 October publishes a drop rate, an offline payout formula or a stage reward table, and the store page’s idle claims are qualitative. Everything quantitative below comes from Valve’s own endpoints, or from Genesis Archive — a player-run wiki that states it reads the game’s content tables from the client, in a snapshot it dates 12 September 2026, four weeks before the launch patch.
What this covers: the drop-table structure behind a stage clear, the synthesis ladder that gates gear tier, the daily caps that bound a session, and the upgrade order the published systems reward. Why it matters: an idle game’s real currency is unattended hours, and the difference between a good and a wasted night is which stage you parked on and which upgrade you queued first. Who should care: anyone already running it who wants to stop guessing, and anyone deciding whether a top-100 free idle game is worth a permanent window on their desktop.
The build this advice assumes
Advice ages badly in a game that patches weekly, so pin it. This describes the live branch as of the 21 October 2026 update: the newest post in the feed, published 20 October for a maintenance window that ran 13:00–15:00 KST the next day. No patch number appears in any of the notes, so the update date is the only identifier the developer gives; the client tables quoted below come from the v1245 snapshot.
That gap matters: the 7 October update added artifact tiers 3–6, made Lv. 105 gear synthesizable and reworked raid rewards, so a table row and your client can disagree — and the client is right.
| Date (2026) | Event | What it changed for a returning player |
|---|---|---|
| 17 Sep | Update | Raid and World Boss entries capped at 10 per day each; alliance ranking defined as the top 100 members’ combined damage |
| 19 Sep | Peak concurrency, 52,835 | SteamDB’s figure, on the game’s own feed |
| 7 Oct | Official launch update | Artifacts (tiers 3–6), Lv. 105 gear into synthesis, G-Coin, Machina rewards for non-first alliances, combat-power board shows the top 15,000 |
| 21 Oct | Update | Auto-Combine ceiling raised from Tier 2 to Tier 5; league promotion for top-50 finishers |
How a stage actually pays out
A stage resolves into three things a player cares about: a flat reward per clear, a per-roll drop chance, and a set of loot pools the roll picks between. The pick is two-stage — choose a pool, then an item inside it — and the wiki transcribes pool weights separately from per-item chances, which is why a “0.5% pool” and a “2% item” are different numbers below.
The timer matters more than any single rate. Every stage sampled carries a 5-second drop cooldown (DropCoolTime): drops are not tied to your clicks or even to a kill, so the engine samples every five seconds for as long as the hero keeps clearing.
| Stage | Gold per clear | EXP per clear | Drop chance per roll (monster table) | Gear pool | Tier 2 pool | Tier 3 pool |
|---|---|---|---|---|---|---|
| Normal 1-1 | 700 | 840 | 12% | 90.0001% | 1.0012% | 0.5006% |
| Nightmare 2-5 | 15,741 | 39,105 | 1% | 74.6675% | 10.17% | 5.085% |
| Doom 4-10 | 93,510 | 356,400 | 1% | 71.5084% | 12.24% | 6.12% |
Two things fall out of that. First, the per-roll drop chance falls as difficulty climbs — 12% at the entrance stage against 1% at both higher samples — while the pools rebalance upward: gear’s share of the monster table drops from 90.0% to 71.5%, and the Tier 3 pool rises from 0.50% to 6.12%. A Tier 4 pool at Normal 1-1 is 12% of a 0.005% share, some six chances in a million per roll; at Doom 4-10 it is roughly one in eighty thousand. Higher stages buy tier quality with drop volume.
Second, the flat reward per clear climbs two orders of magnitude, 700 gold to 93,510, while the drop rate falls. Gold and EXP scale on a different axis from loot. For account level the correct farm is the highest stage you can finish; for a specific tier it becomes a question of pool weights and clear speed, which no published table answers.
The stage index is not a line, either. Genesis Archive renders it as 400 stages: four areas — Ancient Ruins, Tropical Rainforest, Underground Prison, Ice Castle — crossed with ten difficulties from Normal to Doom, ten steps apiece. The FAQ puts both on the same bottom-of-menu panel, so difficulty is a setting, not progress you consume.
The synthesis ladder is the real progression curve
Gear tier does not drop into your inventory at the top. Tier climbs through synthesis, and the client tables price that climb bluntly: 100% success at Tier 1, 90% at Tier 2, 70% at Tier 3, 50% at Tier 4 and 30% at Tier 5 — the same ladder for equipment, accessories, jewels, pets, wings and runes, at six materials per attempt for gear and jewels, three for accessories, four for the rest.
That row explains most of the argument about upgrade order. A Tier 5 attempt fails seven times in ten, and each failure eats the material you fed it, so Tier 5 is a sink you enter only with duplicates you were not going to wear.
Two official notes complicate the arithmetic. The 21 October update raised the Auto-Combine ceiling from Tier 2 to Tier 5 for equipment, accessories and jewels — a September guide saying auto-combine cannot touch high tiers is out of date, and anyone who left it on the old setting fed nothing above Tier 2 until 21 October. The 7 October notice is a caution of another kind: it documents an equipment level-reduction display bug and a Gems of Luck reissue — fixes that corrected what the screen told players without moving the underlying odds. Read the patch note before you trust a tooltip.
Underneath sits a slower ladder the wiki calls Synthesis levels: one track per item family, 100 levels each, feeding dismantling currency — 1% at level 1 to 50% at level 100.
When the game expects you to leave it running
The store copy is explicit and unspecific in the same breath: the hero “sets up camp” in “a small corner of your screen,” there are “no controls required,” and “whether you’re working or sleeping, your hero’s growth never stops.” Steam’s tags read Desktop Companion and Idler before RPG.
Nothing beyond that is published: no offline cap, away-reward rate or catch-up mechanic, and the 5-second drop cooldown is the only cadence the tables expose. What the notes do publish is a set of daily ceilings on everything except the stage farm.
- Raid and World Boss: 10 entries per day each, capped on 17 September — a twenty-entry budget.
- Contract grants are daily, not hourly: Expedition Contract distributes 5,000 Training Tokens per day and Subjugator’s Contract 5,000 Unlock Stones per day, and the same patch made those benefits live in off-seasons.
- Frame currency is capped at 5 units per day from Tier 6 skin cutscenes, regardless of which of the two currencies drops.
Those ceilings describe the intended split. The stage farm is unbounded: leave it open and it ticks on its timer. The parts that pay league placement are metered, so burning an entry while you sleep is a loss; they are worth more in a session you are watching.
The upgrade order that avoids the wall
The wall here is a stage where the boss outlasts you. The wiki’s per-stage benchmark makes the gap visible — 19 required DPS at Normal 1-1 against 87,105 at Doom 4-10 — and the clear condition is killing the boss. When you hit that, gear tier is the lever, and gear tier is gated by three published systems. Spend in this order:
- User level, to 17 and then 19. Auto-Combine unlocks at User Level 17 per the official FAQ, and the wiki’s user-level table puts Auto Synthesis at 17 and the Marketplace at 19 — you cannot sell a drop on the Steam Community Market until the account is levelled, not the character. Account EXP comes from clears, so this is the one goal where “farm the highest stage you can finish” is unambiguously right.
- Training Ground and Ability Unlock. The FAQ is direct: higher levels “increase the probability of higher ranks appearing.” These are odds on a roll, not flat stats, which is why they survive every gear tier.
- Auto-Combine, now that it reaches Tier 5. Set it deliberately. Feeding Tier 3 junk into a 70% roll to chase a Tier 4 you cannot equip is worse than dismantling it.
- Synthesis only on surplus. 50% and 30% are the tiers where a bad night costs a week, and a proportional success bonus does not rescue a 30% base.
- Equipment level-reduction before you chase tier. The client table prices it: a Tier 1 item at 0–1 reduction succeeds 90% of the time for one Whetstone, falling in five-point steps to 45% past ten reduction. Wearing a higher tier early is cheaper than synthesizing one, and that stat needed a hotfix and a Gems of Luck re-issue in October.
- Pets and Wings before a long unattended stretch. The 17 September note fixes a bug where pet EXP gain and wing Gold gain from kills were not applying. Both are kill-reward multipliers, so they compound with the only part of the game that runs on its own.
- Relics last. The FAQ says tier-specific relic materials come from combining or dismantling Tier 3 equipment, accessories, pets, wings and runes, so relic progress is downstream of a decision you make about every Tier 3 item.
The wall is not only a gear problem. Buying a second character of a different class grants stat bonuses through its “Owned Effect” even if you never play it, per the FAQ — but switching class costs the weapon, secondary weapon and skill slots. An alt is a purchase plus a gear set, not a swap.
What is sold, and what the Market actually pays
The game is free with in-app purchases, and Steam files it under Single-player, Multi-player and Online PvP alongside In-App Purchases and Family Sharing. The 7 October update added G-Coin for paid products.
| Offer | Documented terms | What it buys |
|---|---|---|
| Fancy Skin Boxes | Tier 4–6 skins by stated probabilities; skins unlock at user level 20 | Cosmetics, and a daily trickle of frame currency at Tier 6 |
| Roulette | Currency that “replenishes at regular intervals” | Skin boxes and exchange stones |
| Expedition Contract | 5,000 Training Tokens per day during the period | Training Ground progress |
| Subjugator’s Contract | 5,000 Unlock Stones per day during the period | Ability Unlock progress |
| Characters | “Owned Effect” stat bonus per character owned | Account power, independent of play |
| Diamonds (client table) | 3,000 DiaGrowth levels, each adding one stat | Direct stat growth |
The part with no real precedent in a $0 idle game is the exit. Loot is tradable on the Steam Community Market, and the market is not theoretical: browsing the app returns 1,224 listings, from Advanced Jewels at $0.07–$0.09 — 1,074 to 1,253 listings at that price — up to a Special Jewel of Strife at $0.84 and Tier 6 gear at $0.37–$0.60. Cheap tiers are saturated past the point of being worth the click; the spread lives at the top, where the 30% synthesis rate bites. A September notice also announces “strong account suspension measures” against “bot/farm accounts entering the game through abnormal methods,” in the name of a “healthy, free economic system”: a market with a seven-cent floor, backed by bans, is not a side feature.
What the record does not settle
Reception is the number the launch post does not mention. The store page shows 1,960 reviews at 54% positive — Mixed — with 437 reviews in the last 30 days at 38%, Mostly Negative. A large, engaged audience soured inside a month, and none of the 23 posts addresses it: the 21 October note is convenience work and client optimisation. Whatever the complaints are, the record does not name them, and this piece will not guess.
The roadmap is the forward half: Tower of Infinity and character skins in Q4 2026, a 100-player raid in Q1 2027, then a new class, hero and max level cap expansion in Q2 2027. The cap currently reads 100 on the wiki’s user-level table, which is where that Q2 expansion lands.
Three gaps matter, because they are what a player wants and cannot get. No offline-reward formula is published anywhere, so “leave it running” has no rate attached to it. No per-hour farming comparison is published either, which is why the wiki ships a simulator that asks for your own clear time instead of asserting one. And there is no class-balance statement at all, though the 7 October note fixed four skill descriptions — two Ranger, two Wizard — mistranslated in some locales.