set-17deep-diveeconomylevelingrollingintermediate

TL;DR

The "4-cost econ" gameplan is the standard TFT flex archetype: preserve HP through stages 2–3, stack gold to 50 for max interest, power-level to 8 on stage 4-1 or 4-2, and roll down 30–50 gold to 2-star a 4-cost carry plus frontline. The entire plan hinges on a single mechanic: level 8 shop odds triple your 4-cost hit rate compared to level 7 (10% → 30%), so whoever arrives at level 8 first with enough gold to roll hits their spike first. The rest of the plan — stage 2 econ pacing, the "make 10, then 20, then 50" rhythm, health-trading discipline, the 35-gold rolldown floor — all services that single breakpoint.

Key Takeaways

  • Level 8 is the breakpoint. 4-cost odds jump from 10% at level 7 to 30% at level 8 (LoL Wiki TFT Champion page). That's a 3× increase — the single biggest probability jump in the shop odds table.
  • Interest caps at +5 gold (at 50 gold held). The "make 10, then 20, then 50 as quickly as possible" rhythm (Mortdog via Broseph) is the compound-growth engine of the game.
  • Roll down at 4-2 with 30–50 gold. This is the canonical timing. Earlier (4-1) if contested; later (4-3+) loses pool priority to opponents.
  • Minimum rolldown floor: 35 gold (VoidS1n). Below that, you run out of rolls before you hit — the worst outcome.
  • HP is a resource, not a fail state. High-elo players trade HP for interest deliberately, sitting at 40–60 HP at 4-2 to arrive with more gold. "30 HP is a signal to commit, not a signal to stop" (Boosteria).
  • Your stage 2 board is stage 3's foundation. Losing interest thresholds early compounds to stage 5 disasters. Every missed 10-gold breakpoint echoes.
  • Contested 4-costs change everything. Only 10 copies exist per 4-cost unit (pool sizes). Two contestors on one unit is usually fine; three means someone 8ths.
  • Fast-8 isn't universal. Reroll comps, hyper-roll openers, and Set-17-specific fast-9 all have different gameplans. The flex/4-cost plan is the default — know when to deviate.

The shop odds table — the data that drives the whole plan

Every decision in this gameplan is downstream of this table. Shop odds for Set 17, patch 17.1 (current as of April 2026):

Level 1-cost 2-cost 3-cost 4-cost 5-cost
4 55% 30% 15% 0% 0%
5 45% 33% 20% 2% 0%
6 30% 40% 25% 5% 0%
7 19% 30% 40% 10% 1%
8 15% 20% 32% 30% 3%
9 10% 17% 25% 33% 15%
10 5% 10% 20% 40% 25%
11 1% 2% 12% 50% 35%

Sources: League of Legends Wiki — TFT Champion (primary, authoritative), Redeemer TFT shop odds (cross-reference), Riot patch 17.1 notes (confirms the L7 adjustment from Set 16). Some fan sites still show stale Set-16 numbers at L8/L9 — always verify against the Wiki.

What the table actually means

  • Level 7 → 8: 4-cost odds triple (10% → 30%). This is the single biggest probability jump in the entire table — and the mechanical reason fast-8 is the canonical spike.
  • Level 8 → 9: 4-cost odds barely move (30% → 33%). The real value of pushing 9 is 5-cost access (3% → 15%, a 5× jump).
  • Level 6 → 7: 3-cost odds peak at 40%. This is why 3-cost reroll comps sit at level 7 and don't push further.
  • Level 8 is the 3-cost peak at 32%. 3-cost reroll on 7 or 8, never past.

Every 4-cost roll you make at level 7 is 3× less efficient than the same roll at level 8. Getting to level 8 before your rolldown is a massive multiplier on every single gold you spend rolling.

Interest: the compounding engine

The interest formula in TFT is simple and brutal: +1 gold per 10 gold held, capped at +5 gold at 50 gold (LoL Wiki). That cap matters more than it sounds. It means:

  • Holding 50 gold generates +5 per round
  • Holding 100 gold still generates only +5 per round
  • Every gold above 50 is earning zero additional interest

This creates the "50-gold ceiling" — past it, you want to spend gold on levels or rolls, not bank more. Mortdog's canonical econ framing (via the Broseph summary): "Make 10, then 20, then 50 as quickly as you can. After that, spend."

Interest breakpoints matter compoundingly

Here's the subtle point: missing a 10-gold breakpoint on stage 2-5 isn't a 1-gold loss. It's a 1-gold loss per round for the rest of the game until you catch up. Boosteria puts it directly: "a missed 10-gold breakpoint in Stage 2 can echo all the way into Stage 5." If you miss +1 interest for 15 rounds, that's 15 gold, which is an entire level-up or three rolls at your 4-2 spike.

This is the mechanical reason "don't break a streak for a unit that doesn't matter" is a real rule, not just a slogan.

Why 4-2 is the canonical rolldown

Liquipedia defines fast 8 as "going for level 8 as quickly and safely as possible so you can access level 8 roll odds which allow for higher chance of getting 4-cost units" (glossary). The question is when — and the answer almost always lands on round 4-1 or 4-2.

The interest math

Under the standard curve, you'll reach stage 4 sitting on 50+ gold after a few rounds of max interest. At 4-1 (first carousel of stage 4), you have the option to:

  • Level to 8 immediately (costs 4g to buy XP toward the 48 XP needed to reach 8, but if you've been buying a few XP in stage 3, you arrive at 8 around 4-2 naturally)
  • Roll down with the gold remaining

Gamepur's fast-8 guide describes the baseline: "reach level 7 by round 4-1 and power level to 8 as early as round 4-3, with some gold left over to allow you to finally roll down." The faster you arrive with 30+ gold in the bank, the more shops you get at the 22% odds.

The HP tradeoff

The cost of fast-8 is that stage 3 is played under-leveled (you're at level 6 or 7 with a thin 4-unit 2-star board while opponents push trait unlocks at level 7+). You lose more rounds and bleed 10–30 HP across stage 3. The tradeoff is:

  • Fast-8: arrive at 8 first, roll first, 2-star your 4-cost before anyone else
  • Slow-8: arrive at 8 later with more gold, slightly lower hit ceiling but more rolls to convert

VoidS1n's stage 4 guide gives the operational test: "if you have over 35 gold to roll after leveling to level 8/9, especially if you're low HP, you should level and roll." Below 35 gold, you don't have enough rolls to realistically hit — and if you hit below that floor, you've locked yourself into a forced second roll down later with even less cushion.

Stage-by-stage gameplan

This is the standard curve for a flex / 4-cost comp. Numbers are benchmarks, not hard rules — lobby strength, contest, and HP all shift them.

Stage 1 (2-1 through 2-6): the opener

  • Don't full-open. Skip augment combat on 2-1 unless you're clearly going to lose-streak anyway.
  • Buy strong 2-stars whenever they appear — see dpei's econ loop breakdown on why stage 2's five free shops are ~10 gold of value you can't afford to waste.
  • Aim to hit 10 gold by the end of stage 1. This is the first interest breakpoint and the foundation of the "make 10, then 20, then 50" rhythm.
  • Level 4 on 2-1 is the default, level 5 on 2-5 if you're pushing for a stage-2 win streak.

Stage 2 (2-7, the carousel, through 3-1): win-streak or controlled loss

  • Streak. Win-streak if you can; controlled lose-streak if your opener is weak. Both are fine; neither means "sit at 20 gold doing nothing."
  • Target: 20 gold heading into 3-1. Second interest breakpoint.
  • Level 6 on 3-2. This is standard timing. Most flex comps don't push level 7 yet.
  • Start thinking about item direction. Don't slam items you don't need — per dpei's item economy rule, item flexibility at 4-2 is worth more than a premature item slam at 3-2.

Stage 3 (3-3 through 4-1): build toward the spike

  • Level 7 on 3-5 or 4-1. Standard. This is when 4-costs first appear in shops at 10%.
  • Bank toward 50 gold by 4-1. Max interest active.
  • HP target: 60–85 at 4-1. Above this range, you're playing too safe and could have pushed econ more aggressively; below, you're in forced-roll territory on 4-2.
  • Identify your 4-cost target. By 3-5, scout the lobby: which 4-cost carry is most open? Which is contested? Your plan at level 8 depends on which units are realistic to 2-star.
  • Don't slam items yet unless the slam wins you this specific fight. See dpei's strongest-board video on the "if you don't need to slam, don't slam" rule.

Stage 4-1 / 4-2: the rolldown

This is the spike. The entire plan converges here.

  • Level to 8 at 4-1 (if you have buy-XP gold to spare) or 4-2 (more common — you hit 8 naturally from passive XP plus a small buy).

  • Roll with 30–50 gold remaining. Below 35, you don't have enough rolls; above 50, you're over-saving.

  • Priority order when rolling:

    1. 2-star your primary 4-cost carry
    2. 2-star your main tank
    3. Complete utility traits (look for the 2-star of whichever 4-cost closes your key trait breakpoint)
    4. Fill in the 5th and 6th units with whatever appears

  • If contested: roll at 4-1 instead of 4-2. VoidS1n's contested rule: "you would roll most of your gold at 4-1 instead of 4-2 to gain access to these units before your contesters." The extra interest tick isn't worth letting a contestor take all 10 copies of your carry.

  • Slam items. This is the moment — components held from stage 2-3 all convert to completed BiS items now. Critical items: carry's 3 items + 2-3 tank items + any Shred/Sunder/Anti-Heal utility.

Stage 4-3 through 4-7: stabilize and recover

"Stabilize" means: 2-star carry + 2-star frontline + completed carry items + enough HP to survive. Mobalytics defines it as the transition point where you stop panic-rolling and start rebuilding gold.

  • Re-bank toward 30–50 gold for a second spike at 5-1.
  • Position correctly. You're now on your real board — positioning matters disproportionately. Carry in the safe corner, tank up front, CC tank positioned to lock the enemy main tank.
  • Don't push 9 yet unless you have strong board and high HP.

Stage 5-1: fast-9 decision point

This is where the gameplan branches.

  • Push 9 if: HP > 60, 2-star core complete, lobby isn't out-capping you. Level 9 opens the 5-cost pool at 15% (up from 3% at level 8, a 5× jump) — this is the actual value of fast-9, not the marginal 3% gain in 4-cost odds. Game-winners like legendary carries and premier frontline tanks appear here. See dpei's highroll conversion video on why pushing for 5-costs is the correct move once your current board is stable.
  • Stay at 8 and 3-star if: you have specific 4-cost 3-star targets (Level 8 is optimal for 4-cost reroll). Rare for pure flex comps.
  • Don't push 9 if: HP < 40, your board is still missing a 2-star piece. Fix the board first.

HP and gold rolling heuristics

The community has converged on a few numeric heuristics. None of them are absolute rules — they're decision filters.

The 50 HP / 30 HP split

  • Above 50 HP with a playable board: prioritize econ, hit interest breakpoints, delay the rolldown one round if it helps
  • 30–50 HP: this is the committed range. Roll at your canonical spike round (4-2, 5-1) with no hesitation
  • Below 30 HP: you're in late-game forced-roll territory. Spend all your gold, either rolling or leveling-and-rolling. The "30 HP rule" from Boosteria is a signal to commit, not to stop — high-elo players deliberately sit at 30-40 HP while stacking gold, then spike hard

The 35 gold minimum

Across sources, 35 gold is the community-cited minimum for a realistic rolldown at level 8. Below that:

  • You get 3-4 rolls max (at 2g each, 6-8 total)
  • Base shop odds at level 8 give ~1 4-cost per shop (any 4-cost), and you need a specific unit
  • You're likely to bleed out without hitting

If you arrive at your rolldown round with < 35 gold, either skip the rolldown and hold for a 5-1 spike instead, or roll and accept that you're all-in.

The contest multiplier

Contest changes every math above. With 10 copies of each 4-cost in the pool:

  • Uncontested: you have the full 10 copies available. Normal math applies.
  • 1 contestor (someone else 2-starring the same unit): ~5 copies each realistic. Still workable with a standard rolldown.
  • 2+ contestors: pool is effectively emptied. Pivot to a different 4-cost carry or don't force the comp.

Always scout stage 3-5 before committing to the rolldown target. Pivot candidates should be identified by 4-1 at the latest.

Augment interactions

The augment slot at 2-1, 3-2, and 4-2 can amplify or break the plan. General shape:

  • Tempo/econ augments at 2-1 (Two Trick, Heroic Duplicators): straightforward. They give immediate board strength and gold, feeding the econ loop. Fast-8 friendly.
  • Gold augments (Preparation, Golden Ticket, the 2-1 income family): can accelerate the "make 50 gold" curve. Fast-8 staple.
  • XP augments (Level Up, buy-XP econ augments): elite fit for fast-8 and fast-9. Level Up specifically gives a free 2-star at your current cost tier, which is a free 2-star 4-cost at level 8.
  • Combat augments at 4-2 (Infinity Force, Lucky Gloves, hero-specific): fine when your 2-1 and 3-2 were econ. Don't stack three combat augments when you're fast-8 flexing — you'll lack the gold to actually roll.

See live augment tiers at MetaTFT and Mobalytics for current patch specifics.

When fast-8 is the wrong plan

The 4-cost flex gameplan isn't universal. Know when it's wrong:

Reroll comps (1-cost or 3-cost carries)

A reroll comp sits at level 6 or 7 and hyper-rolls down to 0-10 gold, chasing a 3-star low-cost carry. The goals are opposite to fast-8:

  • Level 6-7 is correct (3-cost odds peak at level 7)
  • Spike is earlier (stage 3-5 or 4-1)
  • You want low HP on the rolldown (forcing the commit)
  • Endgame is the 3-star itself, not a 5-cost

Bunnymuffins' leveling guide breaks down the timing: "1-cost rerollers tend to spike at 3-1, 3-5, and 4-1. 2-cost rerollers tend to spike at 3-2, 4-1, and 4-5." Completely different rhythm.

Fast-9 flex

Fast-9 extends the plan: arrive at 8, stabilize lightly, push 9 immediately on 5-1 or even 4-7. Liquipedia defines it as "same as fast 8 but for level 9, usually played when you have a strong level 8 board."

The real value of fast-9 isn't the 4-cost odds — those only go from 30% to 33%. It's 5-cost odds jumping from 3% to 15%, a 5× increase. That's the mechanical reason fast-9 is worth the HP commitment: you suddenly have realistic odds of hitting game-winning 5-costs you couldn't touch at level 8.

The cost: extreme HP commitment. You need > 75 HP at 4-1 and a hittable 2-star level-8 board to make fast-9 realistic. This patch (Set 17.1), several Set 17 overviews flag fast-9 as rough because there's no 5-cost tank to stabilize the frontline — a patch-specific caveat to watch for.

Forced comps

If you've committed to a specific forced comp (see how to force a comp), the gameplan is usually lose-streak control + a heavy 3-5 or 4-2 rolldown on your specific carry — not generic flex-style saving. Forcing overrides the flex gameplan.

Common mistakes

  1. Rolling at 7 for 4-costs. Level 7 has 10% 4-cost odds vs level 8's 30%. Every roll at 7 is 3× less efficient. If you're rolling at 7, you should have a specific 3-cost or tank target — not a 4-cost.
  2. Sitting on 60+ gold at 4-2. Past 50, interest is capped. Every extra 10 gold you hoard is pure waste. Roll or level.
  3. Slamming items on stage 3 for the endgame carry. Your items should serve your current board, per dpei's closest-board rule. The 4-cost endgame carry doesn't exist yet — build for stabilization.
  4. Breaking a streak for a marginal interest tick. +1 gold from hitting a breakpoint is less than +1 win + +1 streak + preserved HP. Streak wins over interest until you're at 50.
  5. Forcing through contested 4-costs. If 2+ opponents are on your carry, the pool is empty. Pivot on 3-5 or 4-1, not 4-3 when you've already burned 30 gold.
  6. Panic-rolling under 30 HP. Random rolls below 30 HP are how you die at 4-5 without spiking. Either commit with a plan (specific units, specific leverage) or hold gold for 5-1. Random rolling kills you.

References and further reading

Mechanics and shop odds

Strategy breakdowns

Mortdog econ philosophy

Augment tiers (live)

Related