Dice engine bug or the game can be hacked

[eroxyde]

I am getting less than 1% odds happening at least twice in my last 4 games + many <3% occurance in my last 4 games. This is not possible at least not to the same player, so there must be a bug or a hack. As for the hack theory, well i ahve also problem with wrong purchase particullary in critical times with Japan, it could be me but i usually verify 2 times and the timing of those errors is kind of suspect !
My winning rate is higher than 80% and i well ranked !
:)
cybernetic war also in the game ?

[eroxyde]

Lol just happenned for the second time in the same game :
Japan : 7 inf, 1 art, 2 ftr, 2 battleships support
vs
Eng : 3 inf, 3 arm

result :
Japan retreat with 1 ftr (this will happens 0.8% over 1000 games)
Eng resist with 1 arm (this will happens 0.7% over 1000 games)

lol that's it for me that game is totally wrong !!!

[Virtus]

Sometimes I wonder myself as I had a 0.5% chance incident happen to me (made a thread about it). Seems more apparent when someone is suicidally aggressive (initiates major battles at less than 40% odds) and somehow consistently wins them.

[MerchantofDoom]

eroxyde の投稿を引用：

Lol just happenned for the second time in the same game :
Japan : 7 inf, 1 art, 2 ftr, 2 battleships support
vs
Eng : 3 inf, 3 arm

result :
Japan retreat with 1 ftr (this will happens 0.8% over 1000 games)
Eng resist with 1 arm (this will happens 0.7% over 1000 games)

lol that's it for me that game is totally wrong !!!

I posted a negative review due to this. I believe some people have found a die script or something of this nature. I first noticed it in a ranked game named uncle canvas. It went 24 rounds because the player went 22-24 or 19-19 round 1 every time I attacked a victory city. I owned everything but the victory cities. Eventually he just didn't have enough cash to deal with my 2-31 or 34s and retreat. Game sucked.
Anyway certain players will roll that 1% chance frequently over multiple games with them. Learn their names and avoid any games with them. Hopefully the devs look into it and I can take down my negative review.

[Bazcat]

Either the online odds calculators are off or something weird is happening. I've lost so many significant 95+% battles (and by significant, I mean the battle was significant, not necessarily large) and often in the same game. And by lost, I don't mean they squeaked out a win at the last roll, but lost by such a margin you'd think I had the 95% backwards.

Other games are perfectly fine and even once in a great while I cringe for the other player's bad luck. However, it happens to me far more than it happens to them.

[ReoHays]

I think anything can be hacked and the online 1st edition in the late 90's, I had the same feeling, but today 2020, I think the developers would be all over this. If they are not, they should be able to look at the data of past games and see how ofton things like this happen and if it's the same people. If they cant do that, maybe they should figure it out.
I get mad at dice, even when I roll them, so i may not be the best judge.

[aardvarkpepper]

@ReoHays:

There's no particular reason developers in 2020 would be all over this. Far as I know, Beamdog didn't have experience developing this type of game. But even if they did, big businesses and governments get hacked all the time.

Far as looking at data of past games - I don't know that Beamdog's got it set up that they can do that. I suggested they put in a game log to .csv output - no response on that - but -

I've written elsewhere I think the developers wanted to target a casual market. That probably means the program was designed and written in a certain way. I don't like it, but it's not unreasonable.

If it was thought that casual players not necessarily with any experience of Axis and Allies would be the best market (and my guess is that *would* be true), why put in features aimed at veteran / more serious players that would cost considerably more to develop? I mean sure, we could say game integrity, player satisfaction, but in the end the decision to develop the game or not was a business call.

And really, if all the features I want had made it into the game, I'm sure it would have taken double the budget - probably more, even a LOT more.

Way I figure it, 1942 Online is in Early Access. My impression is a lot of things are pretty well decided on, but if the community speaks up and says there are issues, then maybe something will be done about those issues. Like with this dice thing.

==

I'll repeat a couple comments here that I've made elsewhere. I don't want these points getting lost, as happens sometimes.

1) The large data sets reported from in-game by players show really bad odds, mathematically. It *could* be that those players were exaggerating, but consider the several hundred hours required for players to carefully document results and compile large data sets to demonstrate validity. As I've written, that's unreasonable to expect of people that aren't getting paid.

2) I don't know that there is a PRNG issue personally. Neither developers nor players have made available data sets that allow data to be filtered and checked for correlations.

ReoHays の投稿を引用：

they should be able to look at the data of past games and see how ofton things like this happen and if it's the same people. If they cant do that, maybe they should figure it out. .

The data I saw the developers release was from some test with undisclosed details that just chunked together a bunch of dice results. But you can't check that for correlations. You need in-game data with full context for that.

So for example suppose I think - for whatever reason - that Russian industrial complex antiaircraft fire hit unusually high, on the first three rounds of play. That level of detail would have to be in the data in the first place if the data were to be checked for correlation.

And as in your quote, if there's a particular player that has suspiciously lucky dice - various filters would have to be set for those as well - for example, percentile results on attacker and defender rolls, respectively, in battles involving more than 30 IPCs of units on each side.

Anyways the program was designed as it was. It would be possible - always possible - to go back and make changes later, but it's usually pretty costly to stick something on afterwards if it wasn't designed to handle it in the first place.

And I'd imagine the developers probably have their hands full with other things as well. So what with one thing and another - I think addressing the question of dice issues is important. But I can't say things like lowering system requirements, improving the UI, and implementing allied carrier use can really take a back seat either.

==

One other note -

If using a tool like AACalc, remember:

1) Make sure the Order Of Losses are set correctly. It's something even veteran users miss.

2) Getting a 1% result is not crazy. Even a 0.1% isn't out of the question. You play how many games with how many battles, and out of those some are going to have odd results. And it may be that you're literally unlucky.

Mind, the player reports of large data sets from in-game I referenced earlier said stuff like what was it - 1 in 27 million, 34 million odds respectively? It's in the sticky thread.

But understand there's a difference between a UNLIKELY (but reasonable-ish) result, and a NEARLY IMPOSSIBLE one. If you're doing a couple hundred battles, you've got to figure you're going to strike out once in a while.

Which is NOT to say players are just overreacting. As I wrote, it MAY be there ARE issues. I'm just saying to think about the mathematics and to realize 0.1% is not as crazy as it sounds.

[ReoHays]

My view is, it is fine, I like what they are doing here. I was just comenting on the post. I can deal with the dice, but I feel the allies are limited, buy not being to use each other carriers and TP's and that tips the advantage to the Axis. Was covered in another post by the developer and you, that the day may come where that will be added to the game, just not today.

[Guderian]

eroxyde の投稿を引用：

Lol just happenned for the second time in the same game :
Japan : 7 inf, 1 art, 2 ftr, 2 battleships support
vs
Eng : 3 inf, 3 arm

result :
Japan retreat with 1 ftr (this will happens 0.8% over 1000 games)
Eng resist with 1 arm (this will happens 0.7% over 1000 games)

lol that's it for me that game is totally wrong !!!

Eroxyde. Sounds like an India scenario. The odds of England surviving is around 1,5% and 4% if they have an AA gun (which you didn't write, but would usually be there in India).

Both scenarios are thus easily within the realm of possible - and does not lead to the conclusion that the game is bugged/hacked.

[jantheman]

Fix it so the game can be enjoyed!

[Kilroy]

aardvarkpepper has comprehensively and usefully laid out the dice simulation issues with the game. I don't have anything to add to the idea that anomalous results could be explained by problems with how dice roles are reproduced in the game. I support making more data available that could answer these concerns

Hacking has not been discussed that much in this forum and, to my mind, is a much more serious threat to the game and the community that supports it. As I remember it, a dice hack destroyed a wonderful A&A community built around the Hasbro game about 20 or so years ago.

I think it would be reassuring if there is a way to report suspicious activity and have action taken if cheating has occurred.

Admittedly, as aardvarkpepper points out, this has to be an economic consideration to the game developers. They can't add a lot of full time staff to police the game for a game that is selling for $20 to what I am sure is a small market.. But, maybe there is a way to crowdsource the review to the community or leverage anti-cheating efforts used by other games.

In any case, these reports are a serious concern to me and I hope they are to the game developers as well.

[tallcupofcoffee]

I strongly suspect all dice rolls and game logic are performed at the server. And the game client is, for all intents and purposes, a dumb client that simply renders board states based on what the server tells it.

This makes hacking unlikely (read, nigh impossible as long as the server has zero trust of the clients).

As to the concerns about dice, I suspect it will be a constant drag on the mainstream viability of the game until a low-luck version is put in. To use Risk as an example, the player can do either true random or low-luck ("balanced blitz" in that developer's parlance) when configuring a game. By far, and it is not even close, most games are configured to "balanced blitz".

[aardvarkpepper]

tallcupofcoffee の投稿を引用：

This makes hacking unlikely (read, nigh impossible as long as the server has zero trust of the clients).

https://en.wikipedia.org/wiki/Cryptanalysis_of_the_Enigma

[zuxius11]

I am definitely going to way in on this one. I've played Axis and Allies for over 30 years and I've seen some crazy things (collectively over 30 years) but this game I just played with Burglar Fork took the cake for all those experiences combined. I believe I did see my three japanese battleships not take their initial hits and instead took them all at the end of combat. My settings were set correctly so that could've been a bug. He then did a poor russian attack on India and won that too with less ground and inferior odds. It was an opportunistic attack and really there was nothing to lose so I'd even consider it to see if I could get lucky. And so he won and that destroyed my very conservative and calculated Japanese game in one turn. From there it seems the game "turned off" in giving me any kind "good luck" whatsoever and it didn't matter if I was defending or attacking. I can write it off as just bad luck but really I've never suffered so many losses back to back in any one game I've played IRL. The final attack in Russia which was under powered (for me) was just me wondering if the game would somehow balance the extreme lack of luck for my side. Instead it gave my opponent massive hits and my forces didn't even meet odds with every roll. I will continue playing ranked a bit longer but this is enough for me to quit ranked and play in closed circles because I don't trust this platform. I've played just two games on this platform and just experienced a game of "upsets for a lifetime." Burger Fork is a great player and the attack on the a Japanese Fleet was one of desperation that turned his game around. Kudos for the victory but if I see something like this in the platform again, I'M DONE!

[Boston_NWO]

The dice are fine. Why is it that the system is rigged when you get unlucky/events don't go your way? This is season 3 and it's the 3rd season where I have over a 90% win rate. I'm not aware of any exploit that I or my opponent has taken advantage of.

[PokerSmile]

I agree. The dice are fine