安装 Steam
登录
|
语言
繁體中文（繁体中文）
日本語（日语）
한국어（韩语）
ไทย（泰语）
български（保加利亚语）
Čeština（捷克语）
Dansk（丹麦语）
Deutsch（德语）
English（英语）
Español-España（西班牙语 - 西班牙）
Español - Latinoamérica（西班牙语 - 拉丁美洲）
Ελληνικά（希腊语）
Français（法语）
Italiano（意大利语）
Bahasa Indonesia（印度尼西亚语）
Magyar（匈牙利语）
Nederlands（荷兰语）
Norsk（挪威语）
Polski（波兰语）
Português（葡萄牙语 - 葡萄牙）
Português-Brasil（葡萄牙语 - 巴西）
Română（罗马尼亚语）
Русский（俄语）
Suomi（芬兰语）
Svenska（瑞典语）
Türkçe（土耳其语）
Tiếng Việt（越南语）
Українська（乌克兰语）
报告翻译问题
讨论规则及指引
1 
举报此帖
Nope. Devs are stuck on the fact that their dice are good as is, regardless of what the community has told them, since DAY 1. Get use to it and " TRY " to enjoy. Good Luck!
This is untrue. We have been testing the system and still continue to test it.
Usually, overall, you will experience both sides in the same game, in the same round even. However, if your 20+ army hits only 1s and 2s when attacking a single unit, nobody will care.
Just fresh from my last round as Japanese. An overwhelming ground force scores only 4/5/6 and I am forced to retreat. Right after this one, a huge balanced fleet battle where I score hits from first round with all my CVs and DDs!!!
And let me tell you a secret. I was 100% sure I will win the naval battle after losing the ground one.
As a suggestion for devs, for big battle - I dunno, anything involving 10+ units for each side - I would use a 2 step random generator. What I mean by this? The numbers are pre randomized, right? Then load 2 sets of numbers for each side, then choose the set that has the minimum deviation. This way, chances to get crazy results should reduce considerably.
(of course, I am a programmer and a math scientist, like everybody else in the forum :)
We don't ignore anything and respond to everything. Please check out the stickied FAQ which contains details on our tests which we did, based on the feedback we had been getting.
one of these commentor's has been rude. hint: it's not the developer!
I think the dice Algorithm is based on IPC disparity, It rewards bad Play with good Luck.
It must also be said, I've seen some dodgy roll's go the other way also vs Myself, This must be the next patch update or we might as well be playing UNO or RISK ?
3 round results:
Lost all attackers and defenders lost 1 infantry. That's beyond frustrating, I would have expected to lose all my infantries, and maybe 1 plane, but when the defending bomber gets a 1 in 2 rounds, something is not right there.
Agreed with others, don't bomb complexes, guaranteed to lose 1 bomber at least each turn.
Maybe this sounds stupid, but can we have it where for example, for every round of a battle, you do a full dice roll for every unit in sequence, then another full dice roll in sequence, then another.., then you just take the average dice rolls for each unit and round to the nearest number. In the players eyes, all they would see is a single dice roll, the multiple rolls per round per unit would be done in the background, and maybe it could be displayed in a log that could be viewed to see what actually happened. Sometimes I want to go back to see the dice rolls when I am on defense to see how i got defeated, especially in huge decisive battles.
Would that stagger it enough to give it more randomness? Maybe that's taking some of that "luck" away from the game, but in my mind, that's not a bad thing. I totally accept when somebody outmaneuvers me and honestly, learn from it. What makes me furious is when the dices are so off that the battle becomes totally lopsided, and it's round 20, and you've poured so much time into it, and the dice algorithm just throws the strategy part of the game away.
Even when I do win these lopsided battles in my favor, I am thinking, I should of lost more units during that battle..
Obviously wouldn't do it that way in the physical board game due to time it takes to calculate, but if it could be done easily, I'd do it that way. Maybe it could be an option to toggle on/off or something.
Putting "fact" in quotes is a fitting way to frame this.
http://calc.axisandallies.org/?mustland=0&abortratio=0&saveunits=0&strafeunits=0&aInf=&aArt=&aArm=&aFig=2&aBom=&aTra=&aSub=&aDes=&aCru=&aCar=&aBat=&adBat=&dInf=1&dArt=&dArm=&dFig=&dBom=&dTra=&dSub=&dDes=&dCru=&dCar=&dBat=&ddBat=&ool_att=Bat-Inf-Art-AArt-Arm-Sub-SSub-Des-Fig-JFig-Cru-Bom-HBom-Car-dBat-Tra&ool_def=Bat-Inf-Art-AArt-Arm-Bom-HBom-Sub-SSub-Des-Car-Cru-Fig-JFig-dBat-Tra&battle=Run&rounds=&reps=10000&luck=pure&ruleset=AA1942&territory=&round=1&pbem=
http://calc.axisandallies.org/?mustland=0&abortratio=0&saveunits=0&strafeunits=0&aInf=&aArt=&aArm=&aFig=1&aBom=&aTra=&aSub=&aDes=&aCru=&aCar=&aBat=&adBat=&dInf=1&dArt=&dArm=&dFig=&dBom=&dTra=&dSub=&dDes=&dCru=&dCar=&dBat=&ddBat=&ool_att=Bat-Inf-Art-AArt-Arm-Sub-SSub-Des-Fig-JFig-Cru-Bom-HBom-Car-dBat-Tra&ool_def=Bat-Inf-Art-AArt-Arm-Bom-HBom-Sub-SSub-Des-Car-Cru-Fig-JFig-dBat-Tra&battle=Run&rounds=&reps=10000&luck=pure&ruleset=AA1942&territory=&round=1&pbem=
It's a dice game, and you need to understand the numbers. Yes, there's reason for legitimate concern over the PRNG, but that *doesn't* mean that *every* deviation from the single most expected result is a problem.
Particularly, a single battle in which the odds weren't that fantastic in the first place isn't legitimate reason to call things ludicrous and talk of lottery tickets. There's perhaps fifty to a hundred combats in any particular game, you should *normally* expect at least some of them not to be "ideal".
That is - the legitimate concern over the PRNG is that there are *too many* such non-ideal battles. But singling out a *particular* non-ideal battle requires really astronomical odds against to be legitimately considered significant evidence of PRNG issues - which the quote doesn't have.
Okay, so you plug in those numbers into AACalc, and get 94.8% win or whatever for two fighters versus one infantry, 50% win or whatever for one fighter versus one infantry. And you think 95% sounds super, so you're going to win etc.
But no. That's not really how it works. And I'll take this opportunity to write about the economics of the attack as well as how players should calculate odds in Axis and Allies.
Two fighters versus one infantry, 1/4 chance both fighters miss, 2/3 chance infantry misses.
1/4 chance both fighters miss, 1/3 chance infantry hits.
3/4 chance both fighters hit at least once, 2/3 chance infantry misses
3/4 chance both fighters hit at least once, 1/3 chance infantry hits.
So 2/12 no hits, 1/12 fighter dies, 6/12 infantry dies, 3/12 fighter and infantry die. But we ignore the first result, because if everything misses,the battle just repeats. And if the attacker thought the attack was a good idea in the first place, nothing chanced, so they have no reason to call off the attack.
Since we're discarding the all-miss outcome as a significant outcome, we now have fewer outcomes - that we're considering, anyways. So the denominator changes, we now have 1/10 fighter dies, 6/10 infantry dies, 3/10 infantry and fighter die. Make sense?
So 1/10 chance utility -10, 6/10 chance utility +3 (the territory can't be captured by air, so only the cost of the infantry is counted), 3/10 utility -7 (lose a fighter -10, kill infantry +3)
As an aside - correctly, a player should understand these numbers do NOT reflect what is a good or bad battle. It ONLY represents the cost of the units involved. Imagine you have a car that costs twenty thousand dollars, it can be rendered inoperative by removing a part worth fifty cents. But the car can still run fine even if a few thousand dollars of nonessential parts are removed. So remember not to focus too hard on the simple costs.
Anyways the costs *do* reflect *some* indicator, if you don't have some tactical reason to carry out the battle *and* it's cost inefficient, then might as well not do it right?
So think on it. 1/10 of -10, 6/10 of +3, 3/10 of -7. Multiply it out and add, and you get -13/10. (-1.3) On a unit count basis, the odds of the battle are favorable. But the expected cost/benefit for the attacker is negative. So it's not a particularly sound attack. (Mind I didn't see the game, and maybe there was good reason for the attack - but absent more information I'd call such an attack "greedy" - that is, it risks too much for a small gain.)
Anyways returning to the topic of the quoted post and the part of my response regarding probability. If using a tool like AACalc, the player gets the impression "95%! that sounds pretty good! let's do it!" But actually as you can see above there's a 1/10 chance a fighter dies yet the infantry survives. Then what happens?
Using similar math to look at 1 fighter versus 1 infantry, we have, ignoring the miss chances, 1/4 fighter misses and infantry hits, 1/4 fighter and infantry both die, 2/4 fighter hits and infantry misses. -11/4 (-2.75). The negative utility is even higher.
So this is where a player needs to understand how AACalc works. Only the overall percentage of attacker winning is output; running the battle to completion assumes battles are to the death. AACalc doesn't know the first casualty results, and folds the whole thing into a single output.
But that is NOT how the tactical evaluation really works. A player doesn't have to "fight battles to the death" and *shouldn't*. Rather, a player should consider the first casualties inflicted and how those change the expected outcomes.
So in AACalc terms, the program "looks" at the battle, figures there's 90% of two fighters killing the one infantry. Then in the 10% case that the two fighters *don't* kill the infantry but a fighter dies, the program runs the new battle of one fighter versus one infantry (which has far worse odds, right around 50% that the attacker escapes unscathed). So you get 90%, plus 50% of 10%, or 95%, which is right about what you see as AACalc's output.
But REALLY, if a player's gone in on that battle, if that 10% result comes up and a fighter dies but not the infantry - SHOULD the player try to go 1 fighter versus 1 infantry? Probably not. The net utility is pretty negative. So maybe the player decides to withdraw if that happens.
What I'm getting at is what looked like a 95% "win" (pretty comfortable) is often thought of as a "comfortable win margin", and players using tools may make decisions on that basis.
But actually the player should understand - especially in this example - that there's a 10% chance of failure, and IF it fails, the unit value of the attackers is very high - and there's no cheap fodder infantry for the attacker to lose as casualties in place of those extremely expensive fighters.
The takeaway is - instead of a "95% win", the takeaway is there was a 10% chance of not just failure, but EXPENSIVE FAILURE. So it was probably a bad risk.
And the other takeaway is 10% isn't likely, but it *is* reasonably likely to happen, especially in a game that has fifty to a hundred battles. And if that 10% happens, so be it, but then if you want to take a 50% battle and are surprised when that doesn't come out, that's gambler's fallacy.
Which I'll note is NOT the same as the INCORRECT distorted dismissal that ALL RNG complaints are "gambler's fallacy" Correctly, gambler's fallacy applies to small case scenarios in which odds are independent of one another.
"The Law of Large Numbers" (it isn't really a law, it's more like really likely guidelines . . . ) has numbers gathering around an average over time. When players legitimately write about RNG concerns, it's not a question of gambler's fallacy, it's a question of the PRNG violating the law of large numbers with disproportionate unlikely results.
Anyways - for this PARTICULAR complaint, real talk, if it was two fighters attacking one infantry - yeah, it's unlikely that an infantry gets a hit before a fighter, but it does happen. 10% isn't crazy odds. And if the attacker presses with one fighter against one infantry, 50% is not crazy odds at all.