OKW troops compared to SU/UKF/USF
Posts: 518
OKW vs UKF :
http://steamcommunity.com/app/231430/discussions/0/1621726179565006830/ (Part I)
http://steamcommunity.com/app/231430/discussions/0/1621726179565658074/ (Part II)
OKW vs SU :
http://steamcommunity.com/app/231430/discussions/0/2595630410189594674/ (Part III)
http://steamcommunity.com/app/231430/discussions/0/2595630410193227906/ (Part IV)
OKW vs USF :
http://steamcommunity.com/app/231430/discussions/0/1693785035806468254/ (Part V)
And before anyone does mention it . I am well aware that those tests may be influenced by RNG and to prevent RNG influenced results I did every test (exept tank test with an clear one) at least twice (if they both hadn't the same or around the same outcome I did it 3 ... 4 ... 5 times (till I could see which case is more likely to happen) . Of course that does not eliminate the RNG factor but it should keep him as small as possible.
Posts: 311
Posts: 3145 | Subs: 2
What is the particular reason as to why you only compared OKW infantry to other Allied infantry and not both Wehr and OKW infantry?
Is it to be used as an argument that the OKW are underperforming or something?
Because more often than not the Wehr infantry have been said to be underperforming because of their small numbers.
And just for the record, I'm not attacking you or anything, I'm just wondering of your intentions with these tests you've conducted, that's all.
Posts: 98
Is it matter of interest tests or are you investigating cost effectiveness in comparison to other factions? Why are you testing certain units match ups and not others?
I have two major criticisms though;
You don't give enough information about each test such as: are the units being microed, did they use abilities, what variation was there between tests, etc.
Second, you don't appear to be taking range into consideration. Various unit have different effective ranges and perform poorly at other ranges, eg what would happen if you performed volks vs IS test at medium and close ranges?
Also how on earth can a firefly beat a KT in a one on one slug match or Cromwell reliably beat an OKW P4?!
Posts: 911
They might again think volks and obers are a bit underpowered and nerf them again.
/shitpost
On topic looks like you skiped obers with lmgs vs bren and double bren.
Also it would have been nice to see single vs double brens vs OKW troops.
Posts: 13496 | Subs: 1
I do applaud and encourage people to do tests however, one needs to be clear about the purpose of the tests in order to define the test conditions.
Actually no one can simply contact a test and then see how one can use those results.
Especially since OP draw no conclusion from his tests.
Posts: 98
Actually no one can simply contact a test and then see how one can use those results.
Especially since OP draw no conclusion from his tests.
Are you suggesting that unit tests are worthless?
All manner of tests are useful to some degree or another, with the accuracy and value of results being dependent on conditions defined.
Tests in a vacuum obviously don't reflect the complex interactions of an actual game but various bits of information will aid your tactical decision making. You might also discover things you never knew as well.
If you referring to testing for sake of unit balance, that's another story and far more complex.
OP did go to the effort to run a whole stack of tests and then post it to the balance forum and it would be nice to know why he went to all that effort. Matter of interest, generate balance discuss, got bored?
In science the why is just as important as the how, as it difficult to come to the right conclusion without asking the right questions.
Posts: 2807 | Subs: 6
Posts: 13496 | Subs: 1
Are you suggesting that unit tests are worthless?
...
No I am saying that one can:
either form a theory and then check its validity with appropriate test
or one can perform test and then see if he can form a theory based on the results.
Both are solid scientific approaches.
In the end of the day ,if you do not like the test contacted, run your own set of the test with the parameters you like so then other can come and criticize your work.
Posts: 98
No I am saying that one can:
either form a theory and then check its validity with appropriate test
or one can perform test and then see if he can form a theory based on the results.
Both are solid scientific approaches.
In the end of the day ,if you do not like the test contacted, run your own set of the test with the parameters you like so then other can come and criticize your work.
Ah I see and yes I agree completely.
Posts: 17914 | Subs: 8
Actually no one can simply contact a test and then see how one can use those results.
Especially since OP draw no conclusion from his tests.
In addition to this, one or two repetitions will never warrant reliable result and unless you repeat the test 50-100 times for each case, you do not have a valid result, but pure RNG error margin.
Posts: 518
I have a stupid question, why not both Axis/German Armies vs the Allies, and not only the OKW?
What is the particular reason as to why you only compared OKW infantry to other Allied infantry and not both Wehr and OKW infantry?
Is it to be used as an argument that the OKW are underperforming or something?
Because more often than not the Wehr infantry have been said to be underperforming because of their small numbers.
And just for the record, I'm not attacking you or anything, I'm just wondering of your intentions with these tests you've conducted, that's all.
Ostheer may be following (if I have the time).And the main reasons for making this comparisions were:
- Countering those "Unit X from OKW is OP nerf plz" threads on steam .
- Showing that some OKW units are underperfoming
- Giving other players some kind of advice which units should be used to counter allied units and which better not
Posts: 518
I do applaud and encourage people to do tests however, one needs to be clear about the purpose of the tests in order to define the test conditions.
Is it matter of interest tests or are you investigating cost effectiveness in comparison to other factions? Why are you testing certain units match ups and not others?
I have two major criticisms though;
You don't give enough information about each test such as: are the units being microed, did they use abilities, what variation was there between tests, etc.
Second, you don't appear to be taking range into consideration. Various unit have different effective ranges and perform poorly at other ranges, eg what would happen if you performed volks vs IS test at medium and close ranges?
Also how on earth can a firefly beat a KT in a one on one slug match or Cromwell reliably beat an OKW P4?!
Units were never microed (except KT vs Firefly (but I wrote "NOTE: Unlike the previous comparison this was a battle in which both tanks were allowed to move" next to the screenshot)and no abilities were used (except when it is stated so (like KT vs Firefly "Sherman Firefly with Rockets")).
"Second, you don't appear to be taking range into consideration."
Some tests were done in different ranges . But you are right most were only done on one range .
Posts: 518
congratz on initiative on this... "project"? but i guess you failed hard with this because you need more testing, just because you simulated one fight or so doesn't mean it will always happen, at least 5 of these fights must happen for proper conclusions, so get back to work!
Like I already said :
I did every test (exept tank test with an clear one) at least twice (if they both hadn't the same or around the same outcome I did it 3 ... 4 ... 5 times (till I could see which case is more likely to happen) . Of course that does not eliminate the RNG factor but it should keep him as small as possible.
Posts: 320
You know what really bothers me? For your "tests" it's like you went out of your way to put allied infantry (UKF, USF) at their more effective ranges.
In fact, as someone in the comments pointed out, he replicated the test of Volks vs Conscripts and had the exact opposite results where volks won at each range which brings said tests into question. RNG is a bitch sometimes.
https://www.youtube.com/watch?time_continue=1&v=YAD8hGK3W5A
Posts: 518
You could mathmatically come to the same conclusion just by looking at riflemen and volks dps. Aka riflemen do more dps mid-close range then volks. Seems you put the volks at a position where they would lose.
You know what really bothers me? For your "tests" it's like you went out of your way to put allied infantry (UKF, USF) at their more effective ranges.
In fact, as someone in the comments pointed out, he replicated the test of Volks vs Conscripts and had the exact opposite results where volks won at each range which brings said tests into question. RNG is a bitch sometimes.
https://www.youtube.com/watch?time_continue=1&v=YAD8hGK3W5A
The range has nothing to do with me choosing the more effective range for the allied troops . The range is more or less random choosed (in some cases just random,in others I wanted to cover every range or I choose the range in which both troops perform the best). Of course there are some "unfair" comparisons like Sturmpioneers vs Royal Engineers on mid range but those were not choosen to make the winning units (in this case the Sturmpionners)
look more stronger than they actually are .
And ofc is RNG a bitch sometimes and that's why I did nearly every test at least 2 times (doesn't eliminate the RNG factor but makes him at least smaller)
Posts: 17914 | Subs: 8
The range is more or less random choosed (in some cases just random,in others I wanted to cover every range or I choose the range in which both troops perform the best).
That alone completely invalidates whatever the point of testing was.
Inconsistent methods without enough repetitions for measurement are never going to present accurate and objective results.
Posts: 2885
The problem with this though is that you would need to spend a lot of time just to statically prove the probabilities and distributions that are given a'priori in the game files. So maybe its just easier to go there and use some maths to calculate the distributions you need.
Posts: 3423 | Subs: 1
Posts: 8154 | Subs: 2
That alone completely invalidates whatever the point of testing was.
Inconsistent methods without enough repetitions for measurement are never going to present accurate and objective results.
It is very good that you have the will to perform such tests. But I think you need some help in this task. First, I would suggest learning some elementary statistics. Then you can form a statistical model and with enough repetitions (10-20 minimum) and right statistical test support your theory.
The problem with this though is that you would need to spend a lot of time just to statically prove the probabilities and distributions that are given a'priori in the game files. So maybe its just easier to go there and use some maths to calculate the distributions you need.
This.
I'll recommend if you want to make any proper analysis/test:
-There's a specific plain map for testing which has range indicators.
-5-15-25-35 More or less the 4 ranges you want to test units against each other. You could go for 0-10-etc but at 10 you are in the borderline range for point blank mechanic.
-A more realistic approach, would be to compare the performance after X amount of time, rather than waiting to see both units wipe each other. Say 30s for all units. Sitting troops or tanks till one of them dies, is not how the game plays.
Livestreams
11 | |||||
23 | |||||
19 | |||||
8 | |||||
2 | |||||
1 |
Ladders Top 10
-
#Steam AliasWL%Streak
- 1.831222.789+37
- 2.34957.860+14
- 3.589215.733+4
- 4.1101614.642+2
- 5.305114.728+1
- 6.916405.693-2
- 7.271108.715+22
- 8.721440.621+3
- 9.1041674.607-2
- 10.17146.788+1
Replay highlight
- cblanco ★
- 보드카 중대
- VonManteuffel
- Heartless Jäger
Board Info
8 posts in the last week
39 posts in the last month
Welcome our newest member, kavyashide
Most online: 2043 users on 29 Oct 2023, 01:04 AM