Chaos Knight performance issues

Spellbound1875#4610

This thread aims to discuss issues with Chaos Knight performance post 2.0 and how it can be addressed.

Briefly, the new Chaos Knight Models are so large it appears they lose effectiveness on the charge as less models make good contact to generate impact damage. Reducing models and increasing combat stats and mass is an effectively solution to this unintended behavior. Potentially just increasing mass could work as well.

I've attached some replays for those interested though note I use the "use any unit as lord" mod found on the workshop and your game will crash if it isn't active (I think the stat mod isn't necessary though it's in the below link if you need it).

Mod is available in what is quickly becoming my wacky test mod folder.
https://drive.google.com/drive/folders/1IH134K2cAgPclFJEMbr5ZBokGqY-JZd1?usp=share_link

A loss of damage is observable in comparison to even far cheaper or weaker cavalry in terms of combat stats and Chaos Knights display unusual patterns of positioning within infantry units.

Examples of damage

Odd positioning

The damage gap does close against cavalry with worse stats as sustained combat continues but this makes using chaos knights awkward as you must leave them in combat significantly longer than other cav to perform cost appropriately and consequently open yourself up to significant damage in return. Chaos Knights also appear to get stuck more easily as they push in deeper to infantry formations to engage more entities.

To address this I wanted to experiment with reducing the Chaos Knight entity count and boosting their stats to compensate. I reduced their entity count to 48, increased their mass by 200, increased their hp to 170 and increased their WS to 58 (same AP ratio for both swords and lances). The numbers were selected through examination of other monstrous cavalry and calculating the numbers necessary to compensate for a 20% loss in entities. No adjustments were made to any other stats.

The results of paired tests showed a notable increase in impact damage and a potential reduction in damage output at 15 seconds as sustained trading continued.

Examples at 5 seconds and 15 seconds.

Mod Test 1

Mod Test 2

No Mod Test 1

No Mod Test 2

Results suggest the approach is successful and that the performance disparity is caused by a reduction in impact damage from poor contact. Additionally the reduction in models lowered the rate of cav entities pushing deeply into infantry and getting stuck which improved usability (seen in replays). There is a potential loss of damage in sustained combat implied by the gap closing at 15 seconds (and it makes sense with more attacks being made but the gain isn't massive).

If I were aiming to tinker with this model more I'd consider a minor boost to CB (10% of sword and lance CB at most) to further compensate for the lost sustained damage as this will be more relevant against large targets. I'd also consider reverting the price change, especially if additional buffs are added. I see increased immediate damage and decreased model loss as significant buffs on their own with the unit performance being comparable to old Chaos Knights.

As a side note with 24 models the Swords of Chaos need substantially higher WS to perform effectively. They currently have both lower entities (33% less!) and much worse WS in comparison to Monstrous cavalry even after accounting for AP. I'd recommend massively boosting their WS (+15 min probably higher) to match their damage to other monstrous cavalry units. That change is included in the mod if you would like to test it but here's a couple screenshots to compare with impact and then every 5 seconds out to 20 seconds. The buff is active for the charge and the test is sloppy in a way which should improve the unmodded knights performance. Apologies the pictures are out of order but you can use the damage values/times to make sense of it.

Mod

No Mod

Mind you I'd crank the WS higher than the +15 I added (~90-95 WS same AP) and increase the Swords of Chaos back to their 1800 price point. Archaon's bodyguard should be on the high end of Cav strength.

As a final point it's possible you could obtain the same result simply by increasing chaos knight mass slightly to increase impact damage. Weapon attacks seem to be working fine after all. I do think the survivability and aesthetic gains of reduced model count as positives but they did take a bit more work.

ThibixMagnus#8300

Nice, based on thests it looks like the most practical solution indeed. Another option - though not incompatible - would be to find a compromise between old horse size and current one, and keep the current model size for Dorghar... but that ship may have sailed. I also wonder if the decreased model number would induce some overlap with the monstruous cavs, but I guess the lack of AP remains the defining feature of chaos knights in any case.

Asamu#6386

So, they didn't get a mass increase to go with their new larger models, which really hurts, because they contact more entities, making it harder to pull through units, and causing them to decelerate faster when impacting infantry on the charge (More contact = faster deceleration = less impact damage).

They have 30 more mass than empire knights, despite being nearly twice the size, which is ridiculous.

Spellbound1875#4610 said:

To address this I wanted to experiment with reducing the Chaos Knight entity count and boosting their stats to compensate. I reduced their entity count to 48, increased their mass by 200, increased their hp to 170 and increased their WS to 58 (same AP ratio for both swords and lances). The numbers were selected through examination of other monstrous cavalry and calculating the numbers necessary to compensate for a 20% loss in entities. No adjustments were made to any other stats.

Looks about right. I was surprised they didn't change the model count and stats of the unit when they gave it the new models.

Spellbound1875#4610

On point one yes that's part of their mobility issue but I do think we see the knights moving much deeper as the back ranks push forward which exacerbates the issue. +200 mass alone I don't think would have made them as nimble, for lack of a better word, on it's own. The fact that the entities pushed in less far and didn't get stuck on each other the same way when moving around made repositioning and cycle charging considerably less awkward. I should probably make the time to test that theory more directly rather than just going on my personal experience though.

On point two my guess is they weren't planning to rework Chaos Knights just provide models the community wanted and as such they didn't consider adjusting stats. Does feel like something that should have got more testing before being shipped but the CoC DLC/IE launch did produce a lot of new assets and plenty of bugfixes so I can see how it made it through.

The_real_FAUST#6885

A good post and I agree.

For a unit that was most certainly feared on TT and should be here they are always underwhelming.

Interesting to look at the mass discrepancies their mass should definitely be higher, and to be honest their unit models are a little tooo massive

Particularly irritating are the swords of chaos. They are in effect a unit of exalted champions and that should be reflected as such.

Loupi#8512

swords of chaos need to lose splash attacks, its crippling them

yst#1879

Hasnt touch this game for ages, did they even fix the leaderboard?

Forget about all these balances tbh, years of neglect brings nothing but disappointment. Such a shame

Spellbound1875#4610

Loupi#8512 said:

swords of chaos need to lose splash attacks, its crippling them

Other monstrous cav units have a similar splash profile so I don't think that explains their poor performance (though now I'm curious to see if removing splash attacks from monstrous cav makes them better or worse into infantry).

Biggest difference between other monstrous cavalry units and the Swords is entity count, followed by mass, then by CB/WS. Just removing splash attacks won't make up those differences so the Swords will still be underperforming. They stat adjustments of some kind to hit the appropriate damage numbers given their entity count.