I don't think math works this way TW

正在查看此主题的用户

Wulfsdottir

Sergeant
Long story short, the probabilities shown in percentage values during persuasion dialogues make no sense whatsoever.

First of all, when you show someone that success probability of something you do is 70%, it means that roughly 7 out of 10 times you must succeed.
It's BS though in case of Bannerlord. I had more than 10 lords gathered in a town (army went in for re-supply) and had those nasty mercenaries to sell. I had three possible dialogue options with 70+ % in first case, 5/14% in second and 55% in third.

Now I had to talk with all of them and only 2 times I got single success with all attempts failing eventually.

Another thing:
What does it mean, when the game shows that you have 5% chance of fail and 14% chance of success? What are the other 71%? Where do they go?

Worse part isthat sometimes even save-scumming doesn't help. I tested it on "bring back the daughter" quest and almost all of my responses had more than 70% of success yes I failed to convince her even after loading the game 20th time.

What do you think about this percentage system and how should it be improved? (In case you think it needs to be improved).
 
Yeah i dont get that whole system plus id rather they mask the probability in words "Highly Likely" "Unlikely" etc
 
I dont think thats the finished system as you even have options in an answer like charm.calculating etc and pressing them doesnt do anything.You can see why ppl who didnt know the history of warband would feel scummed by paying a finished product price and getting back not even the skeleton.Anyways they will add percuassion sometime in the near future 'till 2030 i would guess for sure lol
 
Bruh .dude. Bruh it's ea they're learning maths it's ea dude be patient wtf
The fatness of this trolling is so big, my screen is dripping fat because of this...

Worse part isthat sometimes even save-scumming doesn't help. I tested it on "bring back the daughter" quest and almost all of my responses had more than 70% of success yes I failed to convince her even after loading the game 20th time.
It actually works... I remember that I've savescummed about 14 times before Abagai agreed to marry me. And the probability of success was about 64-80% for each option I chose.
 
To be fair, even with a 70% chance you can fail 18/20 tries, but this should be rare.

Just toss a coin 10 times, it´s 50/50 for heads/tails, but often you will have the result as 20/80 or something like this. Do it like 10.000 times and it will be very very close to 50/50.

Savescumming totally works for me when I try to marry. Most of the time 5 reloads are enough for a success.

I hope they replace this system in the far future. Marriage shouldn´t be dependent on "luck". I prefer the old Warband way and I also didn´t like that.
 
FYI I think the random seed is fixed, so you can savescum over and over and if you do the same actions in the same order, it will get the same result.
If the seed says the 70% roll will fail, it will fail even if you reload and try 1000 times. You have to do something different like pick other options, run around the map and do stuff for a little while, etc.
Some games have an option for resetting the seed after loading to make savescumming easier (like XCOM).
 
Oh god, that hurts :sad: .

So even save scumming is a grind now :grin: .

Save > try > fail > reload > go to another town and do whatever > save > try again > fail > repeat

Nooooooooooooooooooo!
 
I also hate the current system...
Honestly I think that it should be more like being allowed or not a specific dialogue option at all or not if you have a high enough skill. Or having a better or worse outcome in a trade, including quests.

So for the crazy mercenary contract selling, you would manage to sell them for a higher or lower price depending on how high your skill is, but you shouldn't fail at it. If anything, you would need to find the right lord to sell to, as a lord with full army shouldn't be interested, whether a lord with peasants/recruits should be... Same with the marriage cost.

But the actual possibility of marriage shouldn't be tied to a random check..... if they can't think of something better, at least they should bring back the old Warband system, which as said by someone else already, was not exactly my favourite system ever, but it worked and in a way kind of made sense as well since it's supposed to take time and visits to court, not just a couple of random persuasion checks.

I do understand it's an EA, but I hope there will be updates on this, sooner rather than later.
 
Even having high Charm doesn't seem to make a huge difference.

I like that your traits can play a role in the choices you have, but yeah the system still needs tweaking.
 
I hate save scumming, but I hate the current persuasion system even more, so I save scum to deal with it. Then I hate myself for compromising my values. So much hate... I should go and take it out on some looters or something.

I haven't found random seeding to be a problem - repeating the same steps on reload can yield different results without needing to change what you do.
 
To be fair, even with a 70% chance you can fail 18/20 tries, but this should be rare.

Just toss a coin 10 times, it´s 50/50 for heads/tails, but often you will have the result as 20/80 or something like this. Do it like 10.000 times and it will be very very close to 50/50.

Savescumming totally works for me when I try to marry. Most of the time 5 reloads are enough for a success.

I hope they replace this system in the far future. Marriage shouldn´t be dependent on "luck". I prefer the old Warband way and I also didn´t like that.
Agree, give the player the opportunity to influence the outcome.
 
Long story short, the probabilities shown in percentage values during persuasion dialogues make no sense whatsoever.

First of all, when you show someone that success probability of something you do is 70%, it means that roughly 7 out of 10 times you must succeed.
It's BS though in case of Bannerlord. I had more than 10 lords gathered in a town (army went in for re-supply) and had those nasty mercenaries to sell. I had three possible dialogue options with 70+ % in first case, 5/14% in second and 55% in third.

Don't take this the wrong way, but it seems like it is you who have some mathematical misconceptions. 70%, according to Statistics, does not mean you ought to have the desired outcome 7 times for every 10 times tried.
What it means is that, for a very large set of trials, at 70% statiscal percentage chance of success, the results distribution should approach 7 success for every 10 tries.

"7 out of 10 times you must succeed" is dangerously close to Gambler's Fallacy
 
You can only save scum so many times too as after a while the game starts to slow down and freeze. Do it too many times and the game outright crashes.
When you have an 80% chance of success and you critical fail instead at a 9% chance. Multiple times.
 
Don't take this the wrong way, but it seems like it is you who have some mathematical misconceptions. 70%, according to Statistics, does not mean you ought to have the desired outcome 7 times for every 10 times tried.
What it means is that, for a very large set of trials, at 70% statiscal percentage chance of success, the results distribution should approach 7 success for every 10 tries.

"7 out of 10 times you must succeed" is dangerously close to Gambler's Fallacy

Exactly, randomness does not mean evenly distributed. There are formulas you can use to work out how probable a final result actually is, for instance if you flipped 100 coins and got less than around 30 heads you would start to be suspicious, but any closer to the expected result of 50 and you wouldn't be able to conclude anything. But for only 10 tries like in your original post, I wouldnt even be that surprised if you got 0 successes.

That's not to say that bannerlords RNG isn't broken, because in warband it was (I think it was weighted towards smaller numbers, and would give the same numbers unusually many times). No RNG is truly random and it's often better to "fake" randomness to make it less unpredictable, because random sequences repeat results much more than we expect. RNG systems that don't do this can often feel like they aren't random even if they are.
 
Don't take this the wrong way, but it seems like it is you who have some mathematical misconceptions. 70%, according to Statistics, does not mean you ought to have the desired outcome 7 times for every 10 times tried.
What it means is that, for a very large set of trials, at 70% statiscal percentage chance of success, the results distribution should approach 7 success for every 10 tries.

"7 out of 10 times you must succeed" is dangerously close to Gambler's Fallacy
I am familiar with gambler's fallacy and I know that with smaller pools the result can be off of that "7 out of 10" that you get from the larger pool. That's why I said "roughly" But that's clearly not the case here.

Before going to that town I had already visited 3 or 4 other lords who also refused and now take into consideration that 1 lord doesn't count as 1 when we're talking about "success" rate in dialogues. I had 3 tries for each of this lords, so it means 9 to 12 dialogue lines with 2/3 of those lines being way over 50% (70+ and 55+) the third line was that weird 5/14 one (again, where does other 71% go?)

And out of those 12 attempts I got successful maybe once... Then head forward to that town with 10+ lords. And they all have those 3 lines. That's more than 30 persuasion lines 2/3 of which is way over 50%.

And out of sum of those dialogue lines I got success 2 or 3 times. Now, how large the pool must be? It is clear that the percentages are distributed wrongly.

Toss a coin 30 times and if it lands on a one side 28 times you have all the right to suggest it is rigged. Now toss a coin which they TELL you is rigged so YOUR side has 70% probability to land and yet it lands the wrong side 28 times? Then you CAN be sure you were scummed.
 
Exactly, randomness does not mean evenly distributed. There are formulas you can use to work out how probable a final result actually is, for instance if you flipped 100 coins and got less than around 30 heads you would start to be suspicious, but any closer to the expected result of 50 and you wouldn't be able to conclude anything. But for only 10 tries like in your original post, I wouldnt even be that surprised if you got 0 successes.

That's not to say that bannerlords RNG isn't broken, because in warband it was (I think it was weighted towards smaller numbers, and would give the same numbers unusually many times). No RNG is truly random and it's often better to "fake" randomness to make it less unpredictable, because random sequences repeat results much more than we expect. RNG systems that don't do this can often feel like they aren't random even if they are.
Yup. Again. It wasn't 10 tries. It's 30+ tries with 2/3 of those tries claiming to have probability OVER 50% (70 in one case). Responded to that above ^_^
 
后退
顶部 底部