Why I Track Artifacts the Way I Do
Most players don't realize that collecting data on their artifacts is what separates someone who randomly gets decent builds from someone who actually understands their character's optimization path. I spent months pulling for Tartaglia before I started using a dedicated Logbook For Genshin Impact Build Ultimate approach, and the artifacts I was getting were garbage—crit rate below 40%, substats I didn't need, main stats that made no sense for a hydro DPS. Here's the thing nobody tells you: you don't need every piece. You need the right pieces, and tracking them systematically is how you find that out without losing your mind. The core problem most players hit is they either hoard everything (taking up inventory space and mental energy) or they check nothing and end up stuck with whatever random artifact drops from the domain that week.
The Logbook Method Explained
I keep a spreadsheet, not some fancy app. It started as a simple Google Sheet with columns for character name, artifact set, main stat, substats, level, and whether it was kept or scrapped. Over time I refined it. Here's the structure that actually works: Column A: Character name — this should be consistent. I use exact names, not variations, because otherwise your filters break. Column B: Set name — Vermillion Hereafter, Pale Flame, Tenacity of the Millelith, etc. Column C: Slot — Flower, Plume, Sands, Goblet, Circlet. Column D: Main stat — crit rate, crit dmg, atk%, hp%, hydro dmg%, energy recharge, etc. Column E: Substat breakdown — this is where most people stop tracking, but it's the most important part. I list each substat separately with its roll count. The critical insight I picked up after losing track of a 5-piece Pale Flame run on Raiden Shogun: you need to record the total refinement score after each level up, not just whether it has "good substats." An artifact with four perfect substats at level 0 is worthless if it doesn't roll into a decent total score by level 16. That number is what predicts long-term viability.
What to Actually Record (The Filtering Problem)
Here's the edge case that cost me an entire gaming weekend. I was tracking artifacts for Hu Tao, and I had this system going for about three weeks. Then I pulled three different 5-star pieces in one domain session—Vermillion Hereafter, all of them with decent-looking substats. My spreadsheet was already filling up. I didn't know which one to keep, so I kept all three, tracked all three, and six hours later I still hadn't decided. This is a real bottleneck. The logbook becomes useless when you log absolutely everything. The workaround I found: never log an artifact that's clearly bad. If it has energy recharge on a weapon piece, or HP% on a damage-focused build, or crit rate below 35% with no other useful substats, just leave it in the domain box and move on. Only track anything that has a chance of being usable. This cuts your logging work by roughly 60-70% and makes the data actually meaningful when you need it. Specific thresholds I use as a quick filter before logging anything:
Crit Rate main stat: needs at least two useful substats (crit dmg, atk%, pyro dmg%) to log Crit Dmg main stat: needs at least two useful substats (crit rate, atk%, hp%) to log ATK% main stat: always log if substat includes crit or elemental dmg
HP%/DEF%: never log for most DPS characters unless you're building a specific support like Bennett or Zhongli
How to Use the Logbook After You've Built It
Once you have thirty to fifty entries for a given character, the patterns start showing. I found this out accidentally when I had about forty pieces logged for Raiden Shogun. The data told me something my gut didn't want to admit: I was pulling too much for Pyro DMG Bonus goblets when the optimal build actually needed ER goblets for her burst uptime. The spreadsheet didn't lie. Fourteen out of sixteen of my well-rounded goblets were hydro/pyro focused, and none of them served her actual build path. Another pattern I noticed: the same artifact pool tends to produce similar results across sessions. If you keep getting artifacts with the same good substats but wrong main stats, you're not unlucky—you're just running domains for the wrong character setup. The logbook reveals this because you can literally count how many times a specific combination has dropped and cross-reference it with your total pulls. When you're deciding between two artifacts with similar scores, check your log first. If you've logged four pieces with crit rate + crit dmg double substats and only one actually performed in-game, that's data. It means something in your rotation or team composition might be better served by a different stat priority than the raw numbers suggest.
My Preferred Tools and Why Spreadsheets Beat Apps
There are apps like Artifact Logger and Yumemi that claim to do this automatically. I tried them for about two weeks each. The problem isn't the concept—it's the friction. Every time you have to open an app, navigate menus, and manually input data, you stop doing it. A spreadsheet sitting on your desktop takes about three seconds to open and the same to fill out. I know this because I tracked the time. The average artifact entry takes me roughly 20 seconds. The app version, with its navigation overhead, took about 90 seconds. That difference compounds over a month of farming. If you do want an app, I'd recommend something that exports to CSV. That way you can move the data to a spreadsheet later when you want to actually analyze it. The best logbooks I've seen include a simple scoring formula in column F that calculates a composite value based on your character's needs, but honestly, even a basic spreadsheet with conditional formatting does this job adequately. One advanced technique I picked up: add a column for "last kept decision." When you decide to keep or scrap an artifact, note the date and reasoning. After six months of data, you can scroll back and see if your decisions correlate with actual performance improvements. This is how I realized I was consistently overvaluing crit rate over crit dmg on my Xiangling build, which explained why her burst damage felt weaker than it should have been despite having 60% crit rate.
When the Logbook Doesn't Help
I should be honest about the limitations. If you're a casual player who only runs domains maybe twice a week and your main complaint is "I get random artifacts," a logbook won't fix that. The data collection itself requires discipline that most people don't maintain past two weeks. The real value shows up for players who are already farming regularly and want to make informed decisions rather than guessing. Also, the logbook can create analysis paralysis. I've seen players spend more time comparing logged artifacts than actually upgrading their best piece. There's a point where the marginal value of logging another entry drops below zero, and that point is different for everyone. For me, it's around sixty entries per character. After that, the additional data doesn't meaningfully improve decision quality. If you're building a new character from scratch, expect to log at least forty pieces before the patterns become reliable. Before that, you're just collecting noise. I usually set a minimum of twenty entries before I consider my logbook data useful for any build decisions.
The Short Version of What Works
Set up a simple spreadsheet with character, set, slot, main stat, substats, and refinement score columns. Only log artifacts that pass a quick filter—if it's clearly bad, skip it. After forty entries per character, look for patterns in what's dropping versus what you need. Cross-reference your logging decisions with actual performance to catch systematic errors in your judgment. Don't let the logbook become a chore; if you're spending more than fifteen minutes a week on it, you're doing something wrong and should simplify the system. That's the whole approach. It took me about three months of inconsistent tracking to build a working system, and another three months of consistent use before I could rely on it. The return on investment becomes clear once you stop relying on intuition for build decisions and start letting the data guide you instead.
Get the Full Details
