Pershing vs Tiger. Shouldn't Pershing be buffed?

Well, that's the point. Pershing doesn't have superior AI in tests.
https://youtu.be/ZeR8ZjeV_uM
https://youtu.be/fJ0hfawIg2A

If you think I have done something wrong here, you are more than welcome to provide more tests here. After all, sample size is too small. But we can clearly see that AI is even(or worse) to the Tiger.

I just gave many reasons why a stationary test is likely to be highly in favour of MG damage, when real match conditions are likely to be against it. Obviously the latter is impossible to objectively test though.

Better AOE on the main gun has many advantages over relying more on the MGs, that are not included in a stationary test at medium range.

Obviously the Better AOE on the main gun has many advantages over relying more on the MGs, that are not included in a stationary test at medium range.

the advantage is there and noone deni it, but again, simple question, Is that aoe advantage enough to off set all the other disadvantages ? Yes or no ?

I just gave many reasons why a stationary test is likely to be highly in favour of MG damage, when real match conditions are likely to be against it. Obviously the latter is impossible to objectively test though.

Better AOE on the main gun has many advantages over relying more on the MGs, that are not included in a stationary test at medium range.

Tiger doesn't fire MG at all in second test(https://youtu.be/ZeR8ZjeV_uM). Yet it won all 3 cases.

I have tested about 30 tests(including not recorded), and result of the first shot doesn't differ much between Pershing & Tiger. They snipe 1~2 model everytime they shoot.

Can we not agree on that the Pershing itself feels like its in a fine state, just way too expensive for what you get?

Reduce the cost and everyone will be happy. Its role is a panther with AI? Then only make it slightly more expensive than a panther and don't put it on par with real heavies like a tiger.

Isn't its AOE significantly better than the Tiger's?
Serealia gives it an AOE score of 28 vs the Tiger's 21.

The MGs are not that reliable in a real match, as they will lose significant value when moving, with other damage sources (where the AOE instantly killing damaged models will surprise your enemy more), with alpha strikes being more dominant over prolonged stationary engagements, and when enemy infantry vets up or the battlefield becomes light cover. They also have less range (35 vs 45 of the main gun) and no possibility to micro for better results unlike using attack ground with the main gun.

I would personally rather have a reliable and strong main gun over stronger MGs.

The Pershing's AoE is absolutely superior to that of the Tiger, while scatter is a bit worse. I'm not sure how Serealia calculates the AoE score, but if it's something akin to average AoE damage times ROF divided by scatter area then the Pershing should still come out on top. The saving grace for the Tiger is the higher ROF, which as it turns out puts it quite close in terms of performance to its allied peers.

Buffing the MGs would suffer from all the drawbacks you've listed of course, and if AoE adjustments aren't ruled out that would be an even better way to approach things. However, MG damage, as negligible as it may be in the late game, does add a layer of consistency by finishing off wounded models in between shots, which is especially helpful for heavies with their huge AoE damage output but relatively tiny OHK radius. It might also be a much safer option than tinkering with AoE values as the results are easier to extrapolate and also affect single models only.

However, if a slight AoE remodeling is possible I'd probably increase the OHK radius a tiny bit (either by increasing AoE near damage or distance) to make the Pershing a bit more wipey than it is now. The exact amount could be quite easy to determine if the desired % relative increase is decided on.

The Pershing's AoE is absolutely superior to that of the Tiger, while scatter is a bit worse. I'm not sure how Serealia calculates the AoE score, but if it's something akin to average AoE damage times ROF divided by scatter area then the Pershing should still come out on top. The saving grace for the Tiger is the higher ROF, which as it turns out puts it quite close in terms of performance to its allied peers.

Yeah what I felt while doing some test is, ROF difference is much huge than I expected from seeing numbers.

I don't even know what AOE score is, but Faster ROF of Tiger helps deal much more dmg in both AI & AT than Pershing.

The Pershing's AoE is absolutely superior to that of the Tiger, while scatter is a bit worse. I'm not sure how Serealia calculates the AoE score, but if it's something akin to average AoE damage times ROF divided by scatter area then the Pershing should still come out on top. The saving grace for the Tiger is the higher ROF, which as it turns out puts it quite close in terms of performance to its allied peers.

Buffing the MGs would suffer from all the drawbacks you've listed of course, and if AoE adjustments aren't ruled out that would be an even better way to approach things. However, MG damage, as negligible as it may be in the late game, does add a layer of consistency by finishing off wounded models in between shots, which is especially helpful for heavies with their huge AoE damage output but relatively tiny OHK radius. It might also be a much safer option than tinkering with AoE values as the results are easier to extrapolate and also affect single models only.

However, if a slight AoE remodeling is possible I'd probably increase the OHK radius a tiny bit (either by increasing AoE near damage or distance) to make the Pershing a bit more wipey than it is now. The exact amount could be quite easy to determine if the desired % relative increase is decided on.

I think pretty much this.
Both tanks have a very hard time missing completely a shot given their low scatter distance and their max AoE, at least when stationary and shooting at loose formations.
I also find the MG damage quite crucial to finish off models, I would not underestimate their effectiveness. Games are often decided by bleeding your opponent out, and it doesn't help you as much if the enemy squad runs away with half health but (almost) full model count.


EDIT:
One thing with the Pershing though is that its POP needs to go up if it is buffed.
The Pershing (if I see correctly) has 19 POP cost, the Tiger 21. Naturally, the Tiger should be better. A cost reduction might also do it. If we go for the "stronger" Pershing, a tiny ROF buff might be work alternatively.

I'm not sure if the pershing need more AI damage, in its current form what it does need is

a- Price reduction to better reflect its value on field

b- survivability boost, increase hp to tiger level

I'm not sure if the pershing need more AI damage, in its current form what it does need is

a- Price reduction to better reflect its value on field

b- survivability boost, increase hp to tiger level

This is an "either a or b" proposal, I assume? Both would probably too much. But why not just tone down the price and enjoy a better doctrinal AI panther? Make it 550/190 or something and see how it goes in beta.

Okay... I have done just a few tests.

"alpha strike" doesn't seems differ much.

However, Tiger indeed deals much less dmg. and takes longer time to wipe squad in one condition.

Tank must be at move while firing all the time. I think it's probably due to 50% vs 75% moving accuracy.

Well, that's the point. Pershing doesn't have superior AI in tests.
https://youtu.be/ZeR8ZjeV_uM
https://youtu.be/fJ0hfawIg2A

If you think I have done something wrong here, you are more than welcome to provide more tests here. After all, sample size is too small. But we can clearly see that AI is even(or worse) to the Tiger.

Just to take the sample size issue up once more, since I feel it's quite crucial to understand what implications the limited number of repetitions in any test regarding these hugely RNG-based results can have.

I've taken a look in the data log for the simulations posted some days ago in the State of Heavies thread and calculated the T2K values using the results from the first 3, 5, 10, ..., 1000 iterations of the same test (similar to the picture below, but assuming a target size of 0.5 for each model) for the Tiger and Pershing.



If you take a look at how the averages vary even for relatively large sample sets of 100 or more repetitions, it becomes pretty clear why in-game tests, as important they undoubtedly are, can be quite deceiving at times. Though all results point into a certain direction (Pershing being a bit faster than the Tiger), the differences can be rather huge if you take only a handful of tests into consideration.



Again, this is in no way meant to devalue the intent and effort put into conducting the tests in the video. It merely shows that for problems seemingly as simple as "how long does it take tank A and B to nuke the target squad on average", it is next to impossible to perform enough in-game test to get to a robust and reliable conclusion.

Could one not plug the game parameters into a statistical model? Should be possible to model this mathematically without having to run thousands of simulations, no?

well in theory it should be possible to calculate a deterministic outcome, though things get frustratingly complicated if you try to work out the math even for a single entity only under ideal conditions, let alone for a whole squad. from what i've tried so far a simulation using the available in-game parameters is the easiest way to approach this, but it's far from ideal of course. there are other ways to do this, like for example the way Hannibal set up his alpha damage calculations, but each has its own limitations.

Could one not plug the game parameters into a statistical model? Should be possible to model this mathematically without having to run thousands of simulations, no?

Depends. It would at the very least be a huge amount of work.
The first issue is how many shots you allow in the first place. Technically, unless Relic rounds their RNG based values at some point, you'd get almost endless possibilities per shot. For the alpha damage calcs that MMX mentioned, I already restricted this to (simplified) 10.000 shots per square meter of scatter area, resulting in often 100.000 possible outcomes for a single shot. Now raise this to the power of shots you need for your calculation. Unless you use a way more coarse method, you'd end up at easily >10^20 different outcomes. For every possible outcome, then calculate the damage done on single model and squad level and summarize the data. So that's not viable. The other possibility would be to use a more geometry based approach for each model, but then you need to deal with more advanced geometry and a lot of other issues.

(EDIT: The above described method is not purely statistical, but running a purely statistical method is quite complicated and requires a lot of expertise of what you're actually doing.)

Long story short:
Just sitting back and let the sim run for 5-15 min is usually the more viable approach. MMX showed how the data converges after a couple of hundred samples. If this sub-dataset can represent the larger picture well enough, I'd rather work with 1000 data points rather than hundred billion billions.

Depends. It would at the very least be a huge amount of work.
The first issue is how many shots you allow in the first place. Technically, unless Relic rounds their RNG based values at some point, you'd get almost endless possibilities per shot. For the alpha damage calcs that MMX mentioned, I already restricted this to (simplified) 10.000 shots per square meter of scatter area, resulting in often 100.000 possible outcomes for a single shot. Now raise this to the power of shots you need for your calculation. Unless you use a way more coarse method, you'd end up at easily >10^20 different outcomes. For every possible outcome, then calculate the damage done on single model and squad level and summarize the data. So that's not viable. The other possibility would be to use a more geometry based approach for each model, but then you need to deal with more advanced geometry and a lot of other issues.

(EDIT: The above described method is not purely statistical, but running a purely statistical method is quite complicated and requires a lot of expertise of what you're actually doing.)

Long story short:
Just sitting back and let the sim run for 5-15 min is usually the more viable approach. MMX showed how the data converges after a couple of hundred samples. If this sub-dataset can represent the larger picture well enough, I'd rather work with 1000 data points rather than hundred billion billions.

True enough. I work in engineering modeling/simulations, but more with state space systems in physical applications, so I only know of statistical modeling, my expertise is very limited here. I was thinking in terms of modeling the probability distributions and deriving conclusions from those, but as long as we dont have a mathematician doing this for us, your approach is great, I agree! Let's get back to topic, sorry for my offtopic diversion.

What do you think about the price adjustments and positioning the Pershing explicitly as a better AI panther?

What do you think about the price adjustments and positioning the Pershing explicitly as a better AI panther?

Tough nut.
At least make the Pershing cheaper for its price. If we assume population defines combat value and the price should reflect that, it should be somewhat between Panther and Pershing.
Apart from that, I would not compare it with the Panther at all. Pershing needs to be an focused AI unit. First, due to the lack of late game AI in the USF lineup. Second, because they already have the Jackson plus doctrinally the Easy8 and 76mm Sherman as AT focused specialists.
If the Pershing is made stronger, then after thinking about all of that I'd probably go with a tiny ROF buff. It gives it a bit more AI as well as AT, since Pershing (with increased pop), Jackson and a Sherman already eat a lot of population that you can't use anywhere else (plus the 5 pop that you are somewhat forced to take with the major and the ambulance. Although they give value, they make the whole build more rigid).

Sherman is one of the most cost efficient medium tanks.

Still doesn't change the fact that USF only have Sherman and Jackson while Whermacht has Ostwind, Panzer 4, Stug, Brumbar and Panther. 2 tanks vs 5. So by that logic we make Panzer 4 more efficient while removing Ostwind, Stug and Brumbar and making those 3 tanks doctrinal. The point is USF have holes in its loadout while Whermact does not. If M10 was available non doctrinal for the Stug equivalent, while the 105 Sherman Dozer was the Brumbar equivalent was available non doctrinally as well then it would be a fair comparison but it still stands that USF have less choices available non doctrinally. Which is why the Pershing being Doctrinal, on par with a Non Doctrinal Tank (Panther) makes it universally viewed as shit when it doesn't fill in any holes the USF faction has while not being better than its equivalent.

If only Pershing was in commander that had doctrinal smoke anywhere on the map. Wait but it has.

Are you seriously suggesting that the USF Smoke is on par with Panzer Tactician? Obviously trolling at this point for even suggesting that.

Yes these guys are up to no good they are a part a huge conspiracy for axis to win, I wouldn't be surprised if the mod team is also responsible for the spread the corona virus.

Ah yes the same balance team who has created more imbalance than before they started. The same ones who created the Jaeger Light Infantry cancer and didn't have the vision to see that it was broken into the community was in revolt over how broken it was. Also statistically both Whermact and OKW have higher win rates than any allied faction. The numbers speak for themselves. Look at recent tournaments and how British were almost never picked.

Can we return to reality now?

If you took your Axis bias goggles off people would take you more seriously. For every solid thing you post its like 1000 posts of nerfing anything allied related.

Still doesn't change the fact that USF only have Sherman and Jackson while Whermacht has Ostwind, Panzer 4, Stug, Brumbar and Panther. 2 tanks vs 5. So by that logic we make Panzer 4 more efficient while removing Ostwind, Stug and Brumbar and making those 3 tanks doctrinal. The point is USF have holes in its loadout while Whermact does not. If M10 was available non doctrinal for the Stug equivalent, while the 105 Sherman Dozer was the Brumbar equivalent was available non doctrinally as well then it would be a fair comparison but it still stands that USF have less choices available non doctrinally. Which is why the Pershing being Doctrinal, on par with a Non Doctrinal Tank (Panther) makes it universally viewed as shit when it doesn't fill in any holes the USF faction has while not being better than its equivalent.

It does change one fact of your original post:

USF are limited in tank choices and the tanks they do have are lackluster except for the Jackson.

Sherman is not lackluster in any shape or form, it a great Tank and M36 is the best TD.

Are you seriously suggesting that the USF Smoke is on par with Panzer Tactician? Obviously trolling at this point for even suggesting that.

Never made a claim saying USF smoke is on par with panzer tactician, on the contrary I have clearly said Panzer tactician is better:

...
Yes "panzer tactician" is easier to use than "off map smoke barrage" to protect a tank but lest no go crazy about it.
...

so I am not sure why try to misrepresent what I have posted.

Ah yes the same balance team who has created more imbalance than before they started. The same ones who created the Jaeger Light Infantry cancer and didn't have the vision to see that it was broken into the community was in revolt over how broken it was. Also statistically both Whermact and OKW have higher win rates than any allied faction. The numbers speak for themselves. Look at recent tournaments and how British were almost never picked.

You mean the mod team that in their first patch made Penal completely broken and allowed Soviet to dominated the meta for 6 months?

If you took your Axis bias goggles off people would take you more seriously. For every solid thing you post its like 1000 posts of nerfing anything allied related.

And if you stop posting about how people are biased and focused on actual argument and suggestions people might take you more seriously, but I guess if you want to be a forum warrior it is your right.

The tiger is a different tank in a different faction with a different economy. It's not 1=1 balance. The tiger has no capacity for Jackson support not dual BAR support. The tiger doesn't have self healing armour to support it. The Pershing doesn't fact allied TDs.

The Pershing loses in a slugging match with the tiger because it's SUPPOSED TO.
And in the end What the tiger is and can do is frankly irrelevant.
What matters is that the Pershing is balanced in its own faction for its own cost.
Usf tanks are specifically not intended to outlast enemy armour. The entire faction is built around this concept.
The Pershing is basicly a Panther with AI. Not a tiger of any strain.

The Pershing is a meaty tank for USF standards and is priced according. I WOULD like to see some ability that allows its mobility to be highlighted but frankly it's combat potential (especially with CA) is quite good its a reliable tank.
You cannot ignore the power spike that CA brings when balancing the Pershing.

And here we can see the mentality that has created the enviornment and situation for pershing. Balance team members and old timers, RE: axis mains, have decided that pershing is simply never supposed to beat a tiger, end of argument.

You wonder why pershing has never been good? Because the above comment is exactly how balance team sees it. Pershing isnt supposed to beat Kruppstahl!!11! You'll take your neutered """heavy tank""" and like it regardless, because its what you deserve when picking allies

-Balance team and balance team stans