Земи премиум да ги скриеш рекламите
Објави: 6   Посетено од: 121 users
06.11.2016 - 23:29
So I made a video on basically all I could find and my own knowledge about turnblocking here: https://youtu.be/T2S1ikeR_tQ

And it got me thinking. As turnblocking has basically a part of AtWar, we should be able to know more than just the gists of how it works. That is why I am posting this. In my next AtWar video (it will be about turnblocking) if anyone else can find more information about turnblocking (code specifics/ specifics) that would be great. I am planning on doing a second tests in a week or 2, and any information given , their channel will be promoted (both directly and indirectly probably) in that video.

I say this because I think it is very essential that the beginner community is informed and is helped with AtWar; We would lose members otherwise.
----
I'm back: https://switchupcb.com

Вчитување...
Вчитување...
06.11.2016 - 23:34
Just read laos thread, hell, its not that difficult
----

Вчитување...
Вчитување...
07.11.2016 - 01:09
Zero the clown
Корисникот е избришан
Вчитување...
Вчитување...
07.11.2016 - 11:56
Напишано од Tirpitz406, 06.11.2016 at 23:34

Just read laos thread, hell, its not that difficult


lao's guide is inaccurate in regards on how the stack merging theory works, as well as wrong in regards of the Turn Block formula. He never cared to complete or correct the guide anyway, so I wouldn't suggest it as the best thread to learn about Turn Block. More dangerous than that - you will end up with the wrong idea about how Turn Blocks really works.

This video is awesome! Though I don't see it as enthusiast as before due to the current game situation. It also has some inaccuracies - I suggest you to look out and test what you're saying as there is a lot of speculation about how Turn Blocks and priorities works.
Вчитување...
Вчитување...
07.11.2016 - 14:48
Alright so retaking the discussion, so far I know three approach on how the Priorities works in the Turn Blocking formula. The classical approach, the escape approach (which is the one that you put in the video) and the multiplier approach (a new one which I was developing). However all three approach have several flaws. The classical for example, which is the most widespread gives a 100% chances of Turn Block if you have just 1 priority higher than your opponent. Example:

Assume 5 infantries are trying to block 10 infantries that are escaping. At same priority it's 50% chances, however, if the 5 infantries had just 1 priority higher:

Наводници:

50% * 2 = 100%


Aqollo bet for that 100% chances, whereas my experience told me that it was 75% chance. We made several tests and conclude that the formula was indeed wrong (From 14 attempts, 4 failed and 10 success for a ~71% chances of Turn Block). Even if you were to assume that, as some players says, that the max is 99% the test results still contradicts that approach. In definitive, the tests approach more to 75% than to 100% or 99%.






But as I dig more and more info in the forums I found out that we could'be seeing the Turn Block system in the wrong way. According to Ivan:

Напишано од Ivan, 01.03.2015 at 05:02

I don't remember all the specifics of this mechanic, and mostly it was Amok's responsibility. However, what I think happens is that movements by all players are resolved in the same order they were made. So, there's a sort of a queue of movements - they are resolved randomly, but always in the same order. Let's say we have 10 players - so, the first roll is between their 10 first movements. One is selected randomly and executed. Now we 9 first movements and one second movement - again, one is taken randomly. If a particular player is particularly lucky, he might have 2 or 3 of his movements executed, while other players are still at their first movement.

So that's how it works. There's no %, just order and random rolls.


And Amok:

Напишано од Amok, 26.04.2012 at 10:24

Напишано од Guest14502, 26.04.2012 at 10:08

Turns are still prioritized by order. So your last move will just fail to even have that 50% chance of TB.

I should've stated this more clearly in the news update. 50% actually means 100% if your move is a higher priority against the enemy's move. 50% is actually your chance before the moves are processed. So with 5 vs 10 you'll get a 100% chance of turn-blocking if your move gets a higher priority and with 3 vs 10 it will be 60%. Hope this makes sense.


Напишано од Amok, 26.04.2012 at 10:56

Напишано од Leaf, 26.04.2012 at 10:51

So if your TB move is higher priority, then your chance of TB is doubled?

TB can only happen if your move is of higher priority. Then it is decided by the unit count difference.


It seems like the Turn Block system includes yet another mechanic before the main one. My last approach was about the classical approach being true (as it came directly from admins anyway); but also an additional (1 / players) probability for your move to actually happen.

Retaking the situation again, with that in mind. See the following queue:

Player A {5 units attack City Q (Priority: 1)}

Player B {1 unit anywhere(Priority: 1), 10 units leaving city Q(Priority: 2)}

There according to the queue's approach it would'be your 1st move vs his 1st move, both 50% chances that the system process them. Let's assume that he wins that 50% chance, so the queues are now like this:

Player A {5 units attack City Q (Priority: 1)}

Player B {10 units leaving city Q(Priority: 2)}

The probability of each move to execute is 50%. Let's assume that the system takes Player A's move. According to the classical formula, you would have 100% chances of TB. So looking at the result:

Player A {}

Player B {10 units leaving city Q(BLOCKED)}

Since there are 2 players, the probability that your first move happens before your opponent is 50%. This gets lower and lower as there are more players. But in this example, the probability that the system process two of your moves in a row is (1/players) * (1/players), which is (1/2 * 1/2) = 1/4, or 25% chances of escaping. This is traduced on a 75% chances of Turn Block which is what we're looking for.

However not enough details are known about such system. I never got to finish the FAQ that I was making so I let this alternative fall behind. It does seems the most complete though, and it would explain the flaws in the classical formula.
Вчитување...
Вчитување...
07.11.2016 - 16:53
I will mark my guide as incomplete but i am inactive and no longer care to pursue this. Off with you clovis.
----
Вчитување...
Вчитување...
atWar

About Us
Contact

Приватност | Услови за користење | Натписи | Partners

Copyright © 2024 atWar. All rights reserved.

Приклучете ни се на

Сподели