Thundering Seasonal Affix Statistics: Quantifying Risk and Reward

About the Authors

This post was written by Dratnos and Krista, and features insight from Tobo and Roiben. Dratnos and Krista both have top 0.1% M+ titles and cast WoW PvE events, while Tobo and Roiben both compete in M+ events like the MDI and TGP. All four can be found creating content on Twitch, among other places! Thanks to Warcraft Logs for the data!

Introduction

Thundering is the first (and potentially only!) Seasonal Affix of Dragonflight. If you want to know more about how it works, expand the details tab here:
Every 65-70 seconds while in combat, Thundering applies either Mark of Wind or Mark of Lightning, which provide identical benefits of a 30% increase in damage and healing for 15 seconds. Players must come into contact with someone with the opposite mark to remove the debuff, or are otherwise inflicted with Primal Overload which is a 5 second damaging stun

Though you receive a Damage and Healing buff while maintaining the affix, players must be cautious and consider tradeoffs and risk factors with different dungeon, key level, and affix combinations.

For this post we used data to investigate some questions around the risk and reward of the thundering affix, such as “How much extra damage are players getting from Thundering?”, “How much time does a Thundering stun cost the run?”, and “Which specs benefit the most from Thundering?”

Data

We used 30,000 Mythic Plus runs logged on Warcraft Logs that happened after the Thundering tank mark change on Jan 20th but before March 1st. We collected at least 100 runs for every key level between 10 (where Thundering begins) and 23. We also collected some logs from level 24-26 but due to limited logged runs there were less than 100 of each of these levels.

We also made sure to collect data for each of the 38 specs in the game, so at least 400 runs from each spec were included in our sample. The runs were all ones in which the dungeon was completed (i.e. not hearthed out of), and were also all runs where the dungeon was finished within 140% of the timer, the same number needed to earn score from a run.

For each of the reports in our sample, we used Warcraft Logs’ powerful API to collect information about things we cared about in each run. This included basic information like the dungeon’s name, level, and affixes; easy to query variables like the specs of each player involved and the total damage they dealt as well as the number of deaths they had; and some more complicated calculations like the extra damage they gained from the Thundering affix.

Use the filter expression in the Damage Done tab of your log to see how much damage everyone dealt while Thundering was active.

IN RANGE FROM type=”applydebuff” AND ability.id IN (396369, 396364) TO type=”removedebuff” AND ability.id IN (396369, 396364) GROUP BY target ON source END

Since thundering is a 30% damage increase, roughly 23% of the damage (30/130) you see with this filter expression on is thanks to Thundering. Then just compare this to your total damage with the filter expression turned off to see how much of your damage came from Thundering!)

Statistics

Upon averaging the data, we found that across all key levels, Thundering added 2.05% extra damage to the group. We also found that as the key level increases, so does the amount of extra damage that groups get out of Thundering. Though the percentage increase is minimal between level 10 and 20, as players complete keys level 20 and higher, they hold onto their Thundering debuffs for longer – gaining more bonus damage from the affix.

(Healers in Green, Tanks in Blue, Melee in Dark Red, and Ranged in Bright Red)

Each spec benefits differently damage wise from Thundering. In general, Ranged DPS specs tended to do the best, and interestingly, the three top performing specs are all DOT based (Shadow, Balance, and Affliction). Perhaps these specs are gaining the most because they lose the least damage while moving to clear Thundering, as their dots continue to tick, and if a Melee and a Ranged player have to clear with each other it’s often the Melee player that will lose more damage. Healers generally gained less than Tanks and DPS, though notably Discipline and Resto Druid did pretty well comparatively – perhaps because most of their damage comes from DOTs.

Players held on to Thundering for about the same amount of time across our sample, but as the key level gets into the 24+ range, there is a noticeable trend towards higher uptime – even if you factored out the higher keys from which our sample size is smaller.

Deaths

When taking a closer at deaths, we wanted to separate the chance of dying into three groups – the percent of players that die within 2 seconds, 5 seconds, and 10 seconds of being stunned, and sort it by key level. As expected, the higher the key, the more likely players are to die if they are overloaded. You may notice that Thundering appears to become less lethal past level 24, but this is likely due to our sample having a small number of keys at that level, and players doing a better job of clearing in those keys (and potentially hearting out after getting stunned, removing the runs from our data set).

(This is the cumulative chance someone will die after getting Primal Overload stunned, Blue is within the first 2 seconds, Red the first 5, and Yellow the first 10)

When we compare classes, we can see there is quite a difference in which classes get overloaded. Tanks are much more likely to get stunned than other roles, with most of the least stunned specs being Melee DPS, especially those like the Rogue and Warrior specs that have abilities to quickly get to whoever they need to clear with.

Looking at the relationship between getting stunned and timing the key, we can see that each Primal Overload stun tends to cost about 5% of the dungeon’s timer, which holds true across key levels though is more likely to be punishing at the highest levels where the timer is tightest.

Affixes and Dungeons

Because Thundering forces players to stack on each other, it’s safe to assume affixes and dungeons that force players apart will cause a higher amount of overloads, and overload deaths. On average, legacy dungeons like Court of Stars and Shadowmoon Burial grounds seem to be more forgiving with Thundering in general. Players were not only much less likely to get stunned, but were also less likely to die after getting stunned.

This makes sense when you consider the intensity of area denial and ground targeted mechanics in the Dragonflight dungeons – consider Azure Vault’s bosses for instance. Leymor often is a huge Thundering problem if the debuff expiring lines up with the Explosive Brand, which requires you to think several seconds into the future to clear early to avoid disaster. Azureblade has an intermission phase full of swirlies and orbs that can cut you off from the person you need to clear with. Telash Greywing has Frost Bomb and Disintegrate, two mechanics that can also stop you from getting close to someone, and of course Umbrelskul revolves around ground denial and movement restrictions, though notably Thundering does suppress the snare from Oppressive Miasma while it’s active!

When looking at affixes, we notice there is an inverse relationship between the number of Primal Overlands in a given dungeon or affix week, and chance of dying from those overloads within 10 seconds. We can note that with explosive, you have the highest chance of dying if you get stunned, while having the second lowest chance of *being* stunned. We suggest this is because if players accidentally get stunned while explosive bombs are out, they’re significantly more likely to explode and kill members, and perhaps this makes players clear more aggressively to avoid those stuns or perhaps those stuns were catastrophic enough to cause players to hearth out of their keys and thereby remove the run from our data set.

Interviews

We shared some of our findings with two top M+ players, Tobo and Roiben, and asked them some questions about the affix! (Answers edited for clarity and length)

Tobo

How did Thundering work in the MDI setting compared to live keys? Did you plan your routes around Thundering timings? Do you think that’s changed how you look at the affix on live now?

We planned a lot around thundering timings. Having bad timings in something like Nokhud literally slowed us down by 1 minute – and it could happen very easily in a long dungeon and when the timings are so tight and precise. We planned around thundering in Azure Vault to make the 1-phase push on Azureblade. Thundering is very important to plan around in MDI because it is such a massive damage increase and stacking it with other cooldowns saves you a ton of time. I think it made me have a worse outlook on the affix because it is so impactful and feels out of your control at times

What do you think about Thundering as an affix for pugs?

I think it’s very bad for pugs. It requires good communication to optimize properly. Of course you can just yolo a bit but it’s underwhelming when you don’t plan pulls or cooldowns around thundering. The stun is also very tilting for pugs and usually ends in wipes or very bad things happening.

Roiben

Are there any min maxes you personally do around thundering / waiting to clear? Any specifics you might want to get into like per boss or interesting trash pack?

What do you think of Thundering in general? Do you prefer this type of seasonal affix or more route-changing affixes? How do you feel about seasonals going away?

I personally dislike thundering very much. I don’t feel like it has a real “seasonal affix touch” and it’s very very punishing. Failing the affix leads to 2-3 people dying in very high keys in most of the cases which then makes groups not really gain as much from the positive aspect. Playing around it can be really good, but being put in some sort of constant pressure situation until it’s actually cleared ist just… annoying. I guess that means I like route changing affixes more? I’m a big fan of Awakened (BfA S4), Tormented (SL S2), Encrypted (SL S3), Shrouded (SL S4). There were all great affixes with a cool seasonal touch. I’m not sure how to feel about not having seasonal affixes anymore. The affixes I mentioned were a net gain for m+ in my opinion, but then again we had to play the same dungeons for a longer time. Having a bonus affix and learning how to play around it was a nice little challenge, however if the dungeon pool changes every season, having a seasonal affix might not be needed anymore.

Conclusion

Many general assumptions we make about Thundering are likely true – the higher the uptime, the more likely you are to get stunned, and more likely to die. Getting stunned is very costly in terms of how much time is lost on average. We can also see interesting relationships between classes and their damage profiles and stun likelihood. We also see note there is an interesting inverse relationship between key-dependent overloads, and chance of dying within 10 seconds.

We are reluctant to make sweeping claims based off of the analysis in this post! Remember that our data come only from players who log their runs, runs which were completed and not abandoned mid-way through, and runs that finished within 140% of the timer. This means we can’t generalize any of our findings to the broader population. There’s good reason to believe that things like our findings on how punishing Thundering stuns may actually be is underestimated. Similarly, the value of Thundering’s damage buff may be overestimated in this dataset compared to the population as a whole, players who log may also be more likely to try and get the most value out of the affix.