With the Commander Revamp in the works, I’ve decided to revisit a topic that has always been discussed quite controversially in the past – the current state of heavy tanks. There’s been a lot of complaints about how these tanks have lost their punch and how they perform even worse than mediums in the AI role now. Hence, I’ve decided throw my hat in the ring once more in an attempt to bring some numbers to the debate. For those of you who are less inclined to read through the following wall of text I’ve included a TL;DR at the end with an executive summary.
Preface
After receiving nerfs to their AoE profiles in the Winter Balance Patch of 2020, many players found the AI performance of heavies across the board to be somewhat lacking and, as a result, the Tiger, IS-2 and Pershing became much less frequent sightings in all game modes ever since.
For me this came a bit as a surprise, as after running the numbers the results seemed to imply that the overall AI performance loss wasn’t even that great – somewhere in the range of 5-20% depending on distance to target and spacing the entities of the squad (Old comparison here).
However, the method I used back then for comparison was admittedly somewhat flawed, since it didn’t take the contribution of the tank machine guns to the overall DPS and model kills into account. This would, for example, make the performance loss of the Tiger seem much greater than it actually is in-game, simply because the Tiger’s MGs are responsible for a significantly greater fraction of the total damage output than, say, those of the Pershing.
On top of that, my previous assessment was based solely on the time it takes to completely wipe a given squad (T2K). While this yields a pretty good approximation of the performance in a lengthy fight, it isn’t really depicting a realistic scenario to happen in an actual game, as the squad would likely retreat once only one or two low-health models are left.
Another very important factor was also not considered back then; namely the discrepancy between damage dealt to a squad and the actual number of models killed per shot. This seems trivial at first, as more damage would be expected to translate into more entity kills. However, this isn’t always the case and there are several examples where a tank deals more overall DPS to a squad but kills significantly slower (as for example the Cromwell vs. Pz.IV).
To address these shortcomings, I’ve conducted a more extensive set of simulations (details to the spreadsheet/algorithm used here) that not only record the average T2K as the metric for comparison, but also capture the damage dealt and kills scored as a function of time over the whole duration of the simulation. Each result represents the average over 2,000 repetitions of the same setup to eliminate fluctuations due to RNG as much as possible. This should give a clearer picture of the impact of the AoE nerfs and allow the changes to the “alpha-strike” capabilities of the tanks (i.e. the damage and kills after the first shot) to be measured.
Results
1) The Tiger
To begin with, let’s take a quick look how the AoE nerfs in the Winter Balance Patch affected the infamous Tiger tank, and further how the old and new Tiger compare to their smaller cousin, the Pz.IV Ausf. H, in terms of AI performance:
Scatter and AoE profile comparison
As visible from the graph above, the Tiger lost a significant chunk of its near AoE damage and suffered a reduction of the near and mid AoE distance. While this didn’t change the gross AoE damage dealt per shot too much, the effective OHK radius was shortened significantly (from 1.333 to 0.5 m), even below that of the Pz.IV. This means the current version of the Tiger should not only be much worse in knocking out individual models than its pre-WBP alter ego, but also inferior to the Pz.IV… but by how much exactly? This of course depends highly on the circumstances, as squad formation, distance and other modifiers, such as damage reduction (through Vet or heavy cover) and received accuracy (only affecting the DPS of MGs) will have a significant effect on the survivability of infantry squads.
To shed some light onto how these circumstances affect the relative performance of the three tanks, let’s break this down into 4 scenarios, each involving a simulated fight (2,000 repetitions) against a squad with 6 entities and a 0.5 modifier to target size (e.g. as if in yellow cover) under different conditions:
1.1) Distance: 35 m; clumped formation
This scenario emulates the rather common situation of engaging a squad at mid to long range, right at the max range of the tank-mounted MGs. The individual squad members are quite tightly bunched up due to hiding inside a crater or some foliage providing yellow cover, thus increasing the chances of multiple models getting caught in the AoE.
In-game representation of test scenario 1
Graphical representation of results for test scenario 1
Comparing the absolute damage (upper left panel) and damage relative to the live Tiger (middle left panel) dealt over time, it is immediately evident that the live Tiger fares only slightly worse than the pre-WBP version (~ 20% after the 1st and 10% after the 2nd shot, respectively), while dealing significantly more HP damage than the Pz.IV in the same timeframe. However, as established before the raw HP damage is not necessarily the best metric to assess the AI performance since any damage that isn’t converted into a model kill can be conveniently healed away for free after retreating from the engagement.
Hence, the number of model kills scored, especially within the first one or two shots fired, is an important performance benchmark to gauge the “alpha strike” (i.e. 1st shot) capabilities of a tank. Looking at the respective model kill graphs (upper and middle right panels), it becomes clear that both the pre-WBP Tiger and Pz.IV perform much better in this regard than the live Tiger (250% and 90% more model kills with the 1st shot, respectively). This performance gap however closes rather quickly after more shots are fired. In fact, the live Tiger overtakes the Pz.IV with respect to model kills already after the 2nd shot and scores 30-40% more kills from there onward. This is also visible from the median T2K values for the whole 6-men squad (value at Σp > 50% in the lower left panel), where the difference between the pre-WBP and live Tiger is very slim (18.5 s and 21.5 s, respectively), whereas the Pz.IV takes significantly longer to wipe the whole squad on average (35.3 s).
1.2) Distance: 35 m; wide formation
In the next scenario the targeted squad assumes a more spread-out formation typical for units out of cover. This reduces the overall effectiveness of AoE weapons as fewer models will be close enough to the explosion at any given time.
In-game representation of test scenario 2
Graphical representation of results for test scenario 2
Due to the larger distance between the individual models of the squad the raw HP damage dealt per shot is lower and the contribution of the tank MGs to the overall DPS is higher than in scenario 1 (as visible from the steeper slope between shots in the graphs). However, the relative performance differences remain quite similar; the live Tiger deals a bit less damage per shot than the pre-WBP version but much more than the Pz.IV, while the difference in the number of model kills after the first shot is now slightly smaller than before. Still, the Pz.IV beats the live Tiger in terms of alpha strike performance, although less drastically. Overall the edge the Tiger has in terms of AI over the Pz.IV is rather small, which is probably due to the decreased value of its bigger AoE in this case.
1.3) Distance: 10 m; clumped formation
While the majority of firefights between infantry and tanks will play out at mid or long range to avoid getting snared, it is sometimes helpful to close in on a defenseless squad or weapon team to minimize the scatter penalties and maximize the DPS of vehicle-mounted MGs. The next scenario tries to emulate this somewhat by simulating a fight against a clumped squad at 10 m distance.
In-game representation of test scenario 3
Graphical representation of results for test scenario 3
As would be expected, the DPS at a range of 10 m is higher than that at 35 m (scenario 1), albeit most of the DPS increase stems from the greater MG DPS at close range. Once again the alpha strike performance is similar to that in scenario 1 with the old Tiger scoring more than 3 x and the Pz.IV almost twice the model kills with the 1st shell than the live Tiger. However, over the entirety of the fight the performance difference between the pre-WBP and live Tiger is again rather negligible (~5% T2K difference), while the Pz.IV takes about 66% longer to completely wipe the target squad.
1.4) Distance: 35 m; clumped formation with 0.5 DR
The last scenario simulates the effect of heavy cover under otherwise identical conditions to the 1st case. With the reduction in AoE near damage introduced with the Winter Balance Patch 2020 the Tiger now deals a maximum of only 120 damage on scatter hits. This means the live Tiger, unlike the old Tiger and Pz.IV, can no longer one-shot entities behind heavy cover outside of natural hits (i.e. successful accuracy rolls), which should have a noticeable impact on its AI performance.
In-game representation of test scenario 4
Graphical representation of results for test scenario 4
Indeed, both in terms of damage dealt as well as kills scored the live Tiger performs significantly worse than the pre-WBP version and only manages to match the average model kill count of the Pz.IV after the 5th shot fired. The alpha strike performance of the live Tiger is therefore pretty awful even compared to a medium tank of half its cost, which in my opinion is something that needs to be looked at. Over the full length of the firefight this performance gap closes again thanks to the higher average AoE damage stacking up continuously, resulting in a median T2K roughly 30% lower than that of the Pz.IV but still 27% higher than the old Tiger’s 30.5 s.
2) The IS-2
Similar to the Tiger, the mighty IS-2 has seen its AoE near damage and distance reduced after the WPB, almost halving the OHK radius to a mere 0.53 m. Overall the changes were less drastic than for its German counterpart, but the impact on its alpha strike performance should be of similar magnitude. To put things into perspective, the AI strength of the IS-2 – pre and post WBP-revamp – against a clumped squad at 35 m range (as in scenario 1 for the Tiger) is compared to that of the T-34/76 for scale.
Scatter and AoE profile comparison
Graphical representation of results for test scenario 1
As expected, the 122 mm cannons of both IS-2 variants clearly provide more raw damage than the T34’s 76 mm gun by a large margin, while the difference in DPS between the old and current IS-2 is almost negligible (~ 10% less alpha damage). However, as in the Tiger’s case, the number of model kills after the 1st shot went down by almost 60% compared to the pre-WBP IS-2, resulting in the T-34 to snipe about twice as many models with the first shell now. Nevertheless, the median T2K values for both IS-2 incarnations are only about half that of the T-34, meaning in longer engagements the IS-2 remains clearly superior.
3) The Pershing
For the last heavy of the bunch, the Pershing, the AoE changes had a similar impact as for the IS-2. The OHK radius was roughly cut down by half while the raw damage potential remained largely unchanged, suggesting a similar drop in alpha strike capability as for the Tiger and IS-2. In this case, the 75 mm Sherman on HE shells serves as the benchmark for AI comparison against a clumped squad at 35 m.
Scatter and AoE profile comparison
Graphical representation of results for test scenario 1
Looking at the damage dealt per shot the old and revamped Pershing, just like the IS-2, are again less than 10% apart. The edge in DPS the M26 has over the HE Sherman is quite a bit lower than in the Tiger/Pz.IV and IS-2/T-34 case, which shouldn’t be too surprising given the great AoE of the M4’s high explosive shells. However, in terms of alpha strike performance the AoE nerfs cut the average kill count of the Pershing after the 1st shot roughly in half, now being ~ 20% lower than that of the HE-Sherman. Again, the Pershing overtakes the M4 in terms of model kills after the 2nd shot, but the overall difference in performance for long engagements is rather small (median T2K for the HE-Sherman is only about 27% higher).
Conclusion
The AoE profile adjustments made in the Winter Balance Patch clearly achieved their intended goal of severely reducing the alpha strike capabilities of heavy tanks without affecting the AI performance in drawn-out firefights too much. As a result, the Tiger, IS-2 and Pershing now deal a great amount of HP damage without sniping individual models in the first shot, thus giving the opponent ample time to react and retreat a squad if necessary. Compared to the standard medium tanks in their respective factions, all three heavies still perform much better over the course of longer firefights, both dealing substantially more HP damage over time and requiring less time for a complete squad wipe.
However, it can be argued that the reduction in alpha strike capability made heavies much less appealing as a means to inflict MP bleed than the cheaper and much earlier accessible medium tanks that don’t suffer the same inefficiency in model sniping. Furthermore, the reduction of the AoE near damage to 120 significantly worsened this problem for units in heavy cover, against which the revamped heavies are much less efficient than their medium counterparts.
While I don’t think anyone would want to see the AI performance of the affected heavies to go back to their pre-WBP level, I guess it would be fair to argue that the AoE near damage reduction might have been a bit too harsh. Reinstating the ability to OHK models behind green cover would probably not affect the current overall AI performance by much, but make heavies bit more fearful when assaulting entrenched positions. Otherwise, I’d say heavies are in a pretty good spot overall and only very slight adjustments, such as the above, would be needed to let them reconquer a spot in the meta without becoming too good.
What are your thoughts on the matter?
Relic Entertainment
cblanco ★ 
보드카 중대 
VonManteuffel 
Heartless Jäger 
