Decorative background for title

Archon: Disclaimers & FAQ

Release of Archon Build Guides & Tier Lists

The team at Archon, the company behind Warcraft Logs, has been working hard to release our new product's beta, Build Guides and Tier Lists.

We have optimized and tracked down gaps in offerings to collect statistics on Gear, Talents, Enchants, Gems, Consumables, and Trinkets from Warcraft Logs entire store of logs to generate recommendations for character builds based on popular choices. Our extensive sample set allows the data to be filtered down to the different difficulty levels, encounters, and key levels. This article will go through some of the disclaimers, caveats, and things we think you should be aware of while using this product.

We are primarily focusing on Retail World of Warcraft in the initial beta release, and Classic Wrath edition is now also out. Other games are on the horizon! Read more about the exciting releases in our initial Announcement Article and the Wrath Edition.

If you have any feedback for our beta release of Build Guides or Tier Lists, please send it to!

Disclaimers & Caveats

The most important disclaimer to us is an aggregated data set does not necessarily apply directly to your specific situation. You should always sim yourself (Check out Raidbots for retail or WoWSims for Classic) if you have specific questions regarding your situation.

This tool is best used in combination with others. You should find your spec guides for a more complete picture and explanations. We can offer some of the 'what', but to learn the 'why', you should read your spec guides and participate in your Class Discord.

Popularity as a Metric

The main metric we have used to display information on Archon is Popularity. What does popularity mean? It is the most widely used items, consumables, gems, enchants, and talents by the number of parses.

This means that the majority of the time the most popular choice is the best choice - especially for things like enchants, gems, and consumables (where every option is usually equally available to every player). However, particularly for gear, there can be scenarios where the most popular choice isn't the best choice in terms of total player power. For example, near the start of a new season, a Mythic+ trinket that is slightly worse than an end-boss Raid trinket will be more popular because it is easier to obtain. As the season progresses, the Raid trinket will likely overtake the Mythic+ trinket as more players start to loot it.

Make sure to use the filters to find out what the best players are using. For example, even if you raid Heroic, looking at the data for Mythic might help you find improvements in your build choices.

It is also important to emphasize that just because a spec has lower popularity, does not make it unplayable. There are a ton of factors to consider, both how fun a spec is, how useful it is in a raid environment (which includes much more than just throughput), and more.

What do the Numbers Mean?

We use the 95th Percentile (on Warcraft Logs) for the throughput (DPS/HPS) numbers. You can view these on the Statistics Page - 95th Percentile.

A percentile is a mathematical term that describes how a score compares to other scores from the same set. It is commonly expressed as the percentage of values in a set of data scores that fall below a given value. A 95th percentile parse means it is higher (in DPS or HPS) than 95% of total parses. We believe this level of play gives a good idea of the potential throughput of a choice while excluding potential anomalies at the top end.

For the Max Key metric, we show the highest level of key that we have seen timed with that build choice.

For Mythic+ Score, we also use the 95th percentile.

Crafted Stat Items

The information necessary to show stats on the items with Crafted Stats picked by the player is currently not logged by Blizzard. Therefore we only show the tooltip with the Random Stats, and we recommend you check out your spec guide for information on what is advised.

Item Level Disclaimers

Retail has a lot of Mythic+ gear and item level variance, and we show the default tooltip for each item, you can generally assume the highest item level available was used. If you are trying to compare items for your use case, you should sim yourself.

Tier Lists

We show four different tier lists on the site. Throughput (DPS/HPS in Raid), Score (in Mythic+), Popularity (Number of parses for the spec in the last two weeks), and Survivability (based on our Death Statistics).

For Throughput, we use the 95th Percentile parse (correcting for variance by using the lower bound of the 95% Confidence Interval). You can see those statistics here.

For Score, we use the 95th Percentile of Mythic+ Score (correcting for variance by using the lower bound of the 95% Confidence Interval). You can see more score statistics here.

The Survivability metric shown on Archon is the opposite of the Deaths metric shown on Warcraft Logs (survivability = 1 - deaths). Read more in the Death Statistics article.

Below the tier lists, you can see the data set we use to generate the tier lists and browse through those metrics yourself. The data shown in this table for Throughput and Score includes the correction for variance.

Frequently Asked Questions

How is Raid data sampled?

For each encounter, we sample the top 50% (or top 1,000, whichever is bigger) of ranks after eliminating duplicate uploads. We then compute builds for items, gems, enchants, talents, etc based on this subset of the data.

These per-boss top-end datasets are then pooled into a single "All Bosses" dataset, which we use to compute builds for the raid as a whole.

How is "High Keys" data sampled for Mythic+?

To compute builds for High Keys use the top 5% of runs ranked by key level and then score for each spec in each dungeon, with two notes:

  1. We use at least 25 runs for each dungeon, even if that is more than 5% of all runs for that spec.
  2. We do not include keys below level 20 in our build data for High Keys, even if that causes fewer than 25 runs to be present.

These adjustments only come into play for specs with low play rates and dungeons with low popularity.

Why do some Embellishments show as "Unknown"?

Unfortunately, information about which embellishments players are using is not fully logged by the game. The game provides exact information about special items, but patches and linings that are added to other armor are not logged.

The Logs view of embellishments lists these as "Unknown".

Why is there an Armory view of Embellishments?

While information about crafted armor patches and linings is not present in logs, it is present on the Armory.

Armory information is not always reliable because it only shows what a character wore when they logged out—not when they completed a dungeon or raid.

Our Armory view for Embellishments includes only data from the Armory that exactly matches all crafted items (both embellished and un-embellished) the character wore in combat—including enchants and gems on crafted items. Players cannot freely swap out embellishments due to limits on crafting (and if they do, they are likely to wear different items), so we believe that this will be reliable.

The Armory view is not shown by default at this time because we are still refining it.

How do you calculate the Stat Priority shown on the Overview page?

Our Stat Priority is built by examining the stats of a character as reported by the game at the start of boss combat or of a dungeon run.

To track this across the large amount of data that we process for the builds shown on Archon, we place the stats into bins. The size of the bins is game-dependent. In Retail WoW, the bin size is currently 250 points of rating.

The small chart attached to each stat shows the exact bins we used, along with the number of rankings in each bin.

To prioritize the stats, we compute the 95% Confidence Interval for the average rating based on the binned data. If the intervals for two stats overlap, we list them as equal (shown side-by-side). If the lower bound for one stat is higher than the upper bound for another, we put a > between them to indicate that.

The amount of rating shown on each stat is the lower bound of the 95% Confidence Interval used for this calculation.

Contact Us

As always, we love to hear your feedback - send us an email at!

Please feel free to also join us on Discord to share any feedback and suggestions, or to ask any questions, and follow our Twitter for updates!

Remove Ads