TLI ToolkitTorchlight: Infinite companionDownload

Guide

TOWER Sequence

How R&D chips land a sequence on your weapon, the odds per code with best picks, how to pick a chip, and what it costs in Mods.

← All guides

Play one R&D from setup to stop ↓

Also called: Basic R&D

Basic R&D, In-Depth R&D, chips, Mods, Expansion Mod, weapon sequence

Last checked against SS14 Tartarus · 7 min read · updated

In TLI Toolkit: TOWER Sequence calculator (reads the R&D screen)

Play one R&D from setup to stop

  1. Choose the weapon type and Basic or In-Depth mode.
  2. Keep only compatible sequences that help your build.
  3. Set a complete-attempt limit; check client costs and stop at the limit.
Equal-value targets: 3|4|7 or 4|5|6. Offered 3 or 4 → choose shared chip 4; both targets remain possible.
Chip and budget examples
  1. Choose the actual weapon type and Basic or In-Depth mode first. Confirm that every target sequence fits that weapon and that its effect benefits your build. Set the maximum number of complete attempts and read the material count in the client. The offer distribution and material-count formula below remain conditional models, not certified current odds or costs.

  2. Worked chip lesson, 2026-11-15: suppose Basic R&D targets either 3|4|7 or 4|5|6 with equal value. First offer 3 and 4: choose 4, which both codes need. Next offer 3 and 5: either keeps one target alive; choose 3 for the Fire target in this example. Last offer 2 and 7: choose 7 and the collected 4|3|7 matches 3|4|7.

  3. Alternative last offer 1 and 2: neither completes that target. The attempt misses even though the first two choices were sensible. Finish or stop according to the client, but do not expect a material refund. Record the miss and start another complete attempt only if your preset allowance permits it.

  4. For a synthetic budget, if the client asks for 2 Mods and your own current price is 3 FE each, one complete attempt costs 6 FE. A three-attempt limit is 18 FE, with no promise of success. If the client instead asks for 4, revise the cost before starting. This arithmetic is an input example, not a market quote.

Base Mod With equal-value targets and the first offer of 3 and 4, the solver gives 24.5% to finish either sequence after picking 4 and 15.0% after picking 3, under the offer model below. If only Cold Penetration helps your weapon, remove 3|4|7 before ranking the offer. Chip 3 would then leave no valid completion.

The short answer

TOWER Sequence R&D adds one extra line (a sequence) to a non-Legendary weapon or shield. Each round you take one of the chips on offer; if your chips match a sequence’s code in any order, the weapon gets it. With best picks, a Basic R&D for a code of three different chips lands about 10.7% of the time, so plan on around 9 tries on average and 21 to be 90% sure. Codes with repeated chips are rarer, and wanting either of two sequences helps a lot.

The probabilities in this guide use the stated community offer model, not certified current game odds. Check the client’s offers and material count before committing currency.

How R&D works

  • It works on non-Legendary weapons, shields included. A sequence adds its effect without replacing the weapon’s own affixes. A weapon holds one sequence; you can repeat R&D until you get the one you want.
  • Basic R&D: you take 3 chips and can land the Intermediate sequences. In-Depth R&D: you take 4 chips; it is the only way to the Advanced sequences.
  • Chips are numbered 1 to 7. A sequence’s code is a set of chips (Intermediate 3, Advanced 4), and the order you take them in doesn’t matter.
  • Each R&D uses Mods: Base Mods for Basic, and for In-Depth the Expansion Mod of the weapon’s class (Warlock, Vanguard, Sniper or Tank). Stopping an R&D halfway gives nothing back.
  • tlidb lists 111 sequences; which ones a weapon can get depends on its type, and the same code can give a stronger version on a two-handed weapon (+16% instead of +8% Fire Penetration, for example).
What the game doesn’t say: how many chips each round offers. The odds below use the model two community calculators agree on: each round offers different chips, any set equally likely, 2 a round in Basic and 3, 2, 2, 2 in In-Depth. It reproduces their published odds exactly. unconfirmed

The odds by code

The chance a single R&D lands one sequence, with the best pick every round, depends only on the shape of its code:

Odds, costs and assumptions
One target sequence, best picks (computed with the app’s solver)
R&DCode shapeExample codeChance per R&DR&Ds on averageFor 90% sure
Basic R&DThree different chips1|2|610.7%9.421
Basic R&DA pair and one more1|1|56.2%16.136
Basic R&DThree alike1|1|12.3%42.998
In-Depth R&DFour different chips1|4|5|710.4%9.622
In-Depth R&DA pair and two more4|5|5|77.4%13.530
In-Depth R&DTwo pairs5|5|6|64.4%22.551
In-Depth R&DThree alike and one more1|1|1|53.3%30.068
In-Depth R&DFour alike6|6|6|61.0%100.0230

“For 90% sure” is how many R&Ds it takes before nine players in ten have their sequence; one in ten needs more.

Wanting more than one sequence

If two sequences would do, count both as targets: the R&D succeeds on either. For a two-handed weapon that would take Fire Penetration (code 3|4|7) or Cold Penetration (4|5|6), Basic R&D:

Odds, costs and assumptions
TargetChance per R&DR&Ds on averageFor 90% sure
Fire Penetration only10.7%9.421
Fire or Cold Penetration17.0%5.913

The two codes share chip 4. Offered a 3 and a 4 in the first round, the 4 leaves 24.5% and the 3 15.0%: the shared chip keeps both codes alive.

How to pick a chip

  • Take a chip your code still needs. Once you hold a chip that is in none of the codes you want, nothing can land in that R&D.
  • With several targets, take the chip that keeps the most codes alive, the shared one first (the 4 in the example above).
  • A chip of the pair first. For a code with a repeated chip, take one of the repeated chips when it’s offered: on 1|1|5, a first 1 leaves a 15.0% chance and a first 5 only 8.2%. You need that chip twice, so it’s the one that must keep turning up.
  • Weigh by worth when the targets aren’t equal: a sequence worth twice as much to you deserves a riskier line.

What it costs

How many Mods an R&D uses isn’t in the game’s help. A community calculator gives 2 for a one-handed and 4 for a two-handed weapon at level 86, and 10 / 20 at level 100. unconfirmed

Odds, costs and assumptions
Mods, price feed 2026-11-10 (FE each, rounded)
ModForFE
Base ModBasic R&D, every weapon0.18
Expansion Mod – WarlockIn-Depth: wands, rods, scepters, canes, tin staffs, cudgels60
Expansion Mod – VanguardIn-Depth: swords, axes, hammers, claws, daggers42
Expansion Mod – SniperIn-Depth: pistols, bows, crossbows, muskets, fire cannons4.5
Expansion Mod – TankIn-Depth: shields8.6

Under these dated prices and assumed material counts, In-Depth on a level-100 Vanguard two-hander would use 20 × 42 = 840 FE of Mods an attempt, and a four-different code takes 9.6 attempts on average. Check the Mods your client asks for before you plan a budget.

Common mistakes

  • Chasing one rare Advanced code when two common ones would do: the chance per R&D can double.
  • Stopping an R&D halfway because it went wrong: the Mods are spent either way.
  • Planning on the average. One player in ten needs more than the 90% column.

Sources

SourceDateUsed for
TOWER Sequence, tlidb.comread 2026-11-11The help text (rules), every sequence and its code
Two community calculators (火炬计算器, 火炬打造模拟器)2026Chips offered per round (the model) and Mods per R&D; facts only, no code used
Price feed, tlicompanion.com2026-11-10Mod prices
TLI Toolkit’s TOWER Sequence solver2026-11Every chance on this page, computed with best picks

When the recommendation does not fit

Troubleshooting
SymptomCheck and action
A target is missing for the weaponCheck weapon class and R&D tier, rather than chasing a code from another item.
The read offer differs from the screenCorrect the held chips and offered chips before acting. The example does not certify screen-reading accuracy.
All remaining targets are impossibleDo not raise the budget to rescue spent Mods. Reassess only before a new paid attempt.

Next, compare one useful target with several genuinely useful alternatives in the calculator. More targets help only when you would keep those results.

How TLI Toolkit helps

The TOWER Sequence calculator in TLI Toolkit: which chip to take, with the chance of each choice and the expected cost
The TOWER Sequence calculator: which chip to take, with the chance and expected cost of each. Open full-size image (new tab)
  • The TOWER Sequence calculator ranks the chips on offer by your chance of landing any sequence you want (or by worth), for every one of the 111 sequences, with the chance per R&D, the expected R&Ds and FE, and a 90% budget.
  • With Auto-read it reads each R&D round from the game screen (Basic or In-Depth, the chips you hold and the chips on offer) and rings the chip to take and Confirm over the game.
  • Mod prices come from the live feed; the model’s offers per round can be edited if the game changes them.

It’s on the app’s Calculators page. Free for Windows 10 and 11: download TLI Toolkit. More gambles with odds: the Divinity board optimizer.

Try it in the app

It’s part of TLI Toolkit, free for Windows 10 and 11.

FreeWindows 10/11, no installWhat’s new

TLI Toolkit is a Windows PC app: open tlitoolkit.com on your PC to download it. The guides and strategies work on any screen.

Written from the game’s own texts, patch notes and community research. Mechanics change between seasons; if something here is out of date, tell us on GitHub.