Very Likely = 100%, Likely = 0%

[Lord Finnish]

Man this is driving me insane. I've been playing as Peru and I've been allied with USA for most of the time. Yet every time I attack someone and ask USA for help, despite every time having "likely" chance of joining my war, they refuse. Only once I succeed in this.

Calling allies should be like making allies: Just tell me yes or no, without this BS which over and over again lies about what it means. For the hell of it I reloaded no less than seven times and every time this "likely" was not enough for USA to join my war.

What made me ragequit was also that when I was attacking the landlocked, OPM Colombia, UK and France both joined the war against me.... wtf. Colombia is not even in anyone's sphere.

 

[Nzill]

Since when is Colombia landlocked though? Fairly sure they have coast. Panama as well. Which is the reason many Great Powers might be friendly status with Columbia, and due to that intervene in the war agaisnt you.

About the "Likely" thing... Yeah. Sometimes unlikely seems to be likely, and likely unlikely.

 

[Lord Finnish]

Since when is Colombia landlocked though? Fairly sure they have coast. Panama as well. Which is the reason many Great Powers might be friendly status with Columbia, and due to that intervene in the war agaisnt you.

About the "Likely" thing... Yeah. Sometimes unlikely seems to be likely, and likely unlikely.

Thing is, I had waged several wars against Colombia earlier. Venezuela took the northern coast, while I took everything else and liberated Panama. Now I was going to attack the capital region to Conquer them, but apparently London and Paris decided that tally ho mon ami, we are coming from around the world to bail out this third world country from its evil neighbor.

 

[medopu]

I'd also like the Yes/No implemented over this ridiculous Likely/Unlikely/.... system.

 

[grommile]

What made me ragequit was also that when I was attacking the landlocked, OPM Colombia, UK and France both joined the war against me.... wtf. Colombia is not even in anyone's sphere.

They don't have to be sphered. Friendly is enough.

 

[Hootieleece]

I like the Maybe, Likely, Very Likely system.......just because I want some uncertainty to my actions. If War/Diplomacy was ALWAYS certain due to specific formulas. Every Country in the World would fight as soon as it had the power to ensure victory.

War is ALWAYS a gamble......otherwise all the the Anectdotes/Aphorisms of Military History would be meaningless. "No Plan survives contact with the enemy" etc.

 

[Lord Finnish]

I like the Maybe, Likely, Very Likely system.......just because I want some uncertainty to my actions. If War/Diplomacy was ALWAYS certain due to specific formulas. Every Country in the World would fight as soon as it had the power to ensure victory.

War is ALWAYS a gamble......otherwise all the the Anectdotes/Aphorisms of Military History would be meaningless. "No Plan survives contact with the enemy" etc.

I'm all in for uncertainty but the likely-maybe system is clearly broken as can be presented by testing it. If you ever have a Likely country turn you down, no matter how many times you reload it will always be the same. Likely should actually mean Likely for me to accept the system.

 

[unmerged(766520)]

Strange, I saw the UK and USA duke it out throughout the entire 1890s, over Peru attacking Colombia.

 

[unmerged(226525)]

Declare War but don't automatically call the US, instead wait a day and call them via normal diplomacy. Their response will then probably change to very unlikely. For some reason AI does not tell you that they have different algorithms to join AFTER the war is declared.

 

[Lord Finnish]

Declare War but don't automatically call the US, instead wait a day and call them via normal diplomacy. Their response will then probably change to very unlikely. For some reason AI does not tell you that they have different algorithms to join AFTER the war is declared.

Tried that and it didn't.

 

[MannheimCouncil]

There has been an issue since EU3 that the "Likely" statement in everything related to diplomacy was to be taken with a grain of salt. At best it was something like 50/50 chance and some people even assumed they must have switched "Likely" and "Maybe" around. Having said that I've had "Likely" nations join me in a war in V2. I'd say the chance can't be much lower than 50/50. I'd agree though that "Very Likely should be around 85-100% and Likely something like 65%-85%. Everything below that should be "Maybe" or worse.

 

[Lord Finnish]

There has been an issue since EU3 that the "Likely" statement in everything related to diplomacy was to be taken with a grain of salt. At best it was something like 50/50 chance and some people even assumed they must have switched "Likely" and "Maybe" around. Having said that I've had "Likely" nations join me in a war in V2. I'd say the chance can't be much lower than 50/50. I'd agree though that "Very Likely should be around 85-100% and Likely something like 65%-85%. Everything below that should be "Maybe" or worse.

Part of the problem is that it seems to be a static figure instead of a truly random chance. CK2 has this as well.

One day, in CK2, I was going to assassinate someone. I had 44% chance to succeed and 10% chance of being detected. I failed the assassination, but also was detected. Okay, just bad luck. But then I tried again. Same result. And again. Same result. Only the sixth or seventh time it was that the assassination succeeded. From that I knew that the random chances lie.

 

[unmerged(46341)]

Part of the problem is that it seems to be a static figure instead of a truly random chance. CK2 has this as well.

One day, in CK2, I was going to assassinate someone. I had 44% chance to succeed and 10% chance of being detected. I failed the assassination, but also was detected. Okay, just bad luck. But then I tried again. Same result. And again. Same result. Only the sixth or seventh time it was that the assassination succeeded. From that I knew that the random chances lie.

That's not proof of anything.

 

[Jorlaan]

That's not proof of anything.

It is and it isn't. It is one example of what does seem to be a problem in these games. I have also been faced with the CK2 problem MANY times, no matter what the % chance is to succeed, if I fail it first time, I'll fail it every time. If it was TRUELY 50/50, then getting 7 no's in a row on a VERY REGULAR BASIS should be so far outside the realm of statistical probability that this discussion should never even be happening.

This is not however limited to Paradox games.

Try any Total War game. If you pit an army up against another and use the "automatic battle" button, there is of course a chance of losing. The balance of the battle is shown in a bar, so you get an idea if its going to work or not. More often than not even with a 50/50 chance you'll win most of the time. However if you lose, and reload, you'll lose again, and again and again and again ad infinite. I have tested this to DEATH and it seems to hold true.

Somewhere, somehow, for some unknown reason, AI's are getting "stuck" in a deterministic outcome of the probabilities. I see this in other games too that use a degree of chance. I have no explanation for this, I am no programmer, I know nothing of AI mechanics beyond what I can observe for myself, and have been observing for a long time. I do however hold this to be true. One thing I have dealt with a lot in my life, is probability. As I have said, the odds of these kinds of things happening on so frequent a basis is so goddamned remote that it should virtually never happen.

Yet here we are.

 

[dz_duck]

It is and it isn't. It is one example of what does seem to be a problem in these games. I have also been faced with the CK2 problem MANY times, no matter what the % chance is to succeed, if I fail it first time, I'll fail it every time. If it was TRUELY 50/50, then getting 7 no's in a row on a VERY REGULAR BASIS should be so far outside the realm of statistical probability that this discussion should never even be happening.

This is not however limited to Paradox games.

Try any Total War game. If you pit an army up against another and use the "automatic battle" button, there is of course a chance of losing. The balance of the battle is shown in a bar, so you get an idea if its going to work or not. More often than not even with a 50/50 chance you'll win most of the time. However if you lose, and reload, you'll lose again, and again and again and again ad infinite. I have tested this to DEATH and it seems to hold true.

Somewhere, somehow, for some unknown reason, AI's are getting "stuck" in a deterministic outcome of the probabilities. I see this in other games too that use a degree of chance. I have no explanation for this, I am no programmer, I know nothing of AI mechanics beyond what I can observe for myself, and have been observing for a long time. I do however hold this to be true. One thing I have dealt with a lot in my life, is probability. As I have said, the odds of these kinds of things happening on so frequent a basis is so goddamned remote that it should virtually never happen.

Yet here we are.

Is the state of the random number generator saved in the save game? That would explain it, if you're just loading the same random number stream every time.

 

[MannheimCouncil]

I'm pretty sure it is not saved in the savegame. I have cheated successfully by reloading. I don't remember whether that was at the "Very Likely" or the "Likely" level, though. I'm not sure whether I progressed another day or so. You might want to try that to make sure the RNG isn't tied to a particular in-game variable.

The likely cause, however, is that "Likely" is used for chances that are not at all what a regular person would regard as likely. With a 40% base chance, the probability that something does not work is over a third even with one reload. And you can pretty much say that sample is biased, because the times it works on the first try you don't really notice it. Not sure how seven reloads in a row can be explained (except as an outlier), but if we want to test it we'd need more test cases. I doubt it's worth the hassle, though. This seems to be too deeply ingrained in the Clausewitz engine for PI to bother to change it.

 

[zyphial]

In my experience, Very Likely is never more than 40%, Likely is no more than 20% and maybe is anything less than Likely. It's absurd, needless obfuscation of very important information. It has nothing to do with certainty, it's a patent lie. 92.6% is an uncertain figure, because while you know it's likely, it isn't 100% or 0% and therefore can go either way. "Very likely" is nonsense.

Load up as the papal states, ally France, justify a war against tunisia and try to get them to join. Nothing I can do will get them to join (ie, 0% chance) and yet it's "very likely." It makes the information completely useless when it doesn't mean what it says it means. Very irritating, and part of the reason I really don't enjoy Vicky 2. Try as I might, I keep finding myself regretting purchasing it because of absolutely inexcusable and downright stupid things like this.

 

[mustachewarfare]

The only thing I really hate is that when i ask for help and gets rejected, the alliance breaks off and some alliance I create it just to maintain good relationship with the targeted nation. So everytime I start a war instead of calling entire allies. I have to painfully manually call individual nations who are willing to help me. This is major pain in the ass and it also means that i need to wait diplomat cool down time which is major hassle for I desire to gain control of my pupet's unit as soon as possible.

 

[Beagá]

In my experience, Very Likely is never more than 40%, Likely is no more than 20% and maybe is anything less than Likely. It's absurd, needless obfuscation of very important information. It has nothing to do with certainty, it's a patent lie. 92.6% is an uncertain figure, because while you know it's likely, it isn't 100% or 0% and therefore can go either way. "Very likely" is nonsense.

Load up as the papal states, ally France, justify a war against tunisia and try to get them to join. Nothing I can do will get them to join (ie, 0% chance) and yet it's "very likely." It makes the information completely useless when it doesn't mean what it says it means. Very irritating, and part of the reason I really don't enjoy Vicky 2. Try as I might, I keep finding myself regretting purchasing it because of absolutely inexcusable and downright stupid things like this.

Duuno if it´s vanilla Victoria 2 you´re playing or with the expansions, because if it´s the expansion what you´re saying is false.

 

[podcat]

Randomness is by definition random, so just because the guy who gets a bad roll 5 times in a row of 10000 people once posts on the forum that doesnt mean those result numbers actually prove anything.

I am pretty sure we save the random generator state as well, although things can happen that change it that you might not notice as the player.

Here are the rules for HoD:
0% Impossible
1-20% Very Unlikely
20-40% Unlikely
40-60% Maybe
60-80% Likely
80-100% Very Likely

This varies a bit between games. I think HoI3 has Likely starting from 70%. This is probably better from a player expectation point of view. Another way is to do weighted random rolls that even out even in small series to avoid people feeling the bad luck, but its not something we have ever tried and I'm not sure its something people want