DeletedUser
I have seen 5 of the same thing be the solution exactly once, so I know it's possible, just not bloody likely.
-
We are looking for you! Always wanted to join our Supporting Team? We are looking for enthusiastic moderators!
Take a look at our recruitement page for more information and how you can apply:
DeletedUser29726
As to the Monty Hall problem reference, the difference is the independance of the queries. You choose to query a door and then based on that the host chooses to reveal a door - his options being constrained by your door choice (in the likely event you're wrong he has to show you the other wrong door, and therefore by switching you become likely right).
In our negotiation game, the answer to every query is independant - so there's no hidden biasing regarding things you did not ask about.
Some other tasks that might be of interest:
Now one thing that would be worth working on is ideal arrangements for partitioning the results with your 2nd last guess. I'm not 100% convinced that the obvious tries there are the best tries to optimise probability. I've done well enough so I haven't really approached the problem yet beyond what i could work out in my head in the 5 seconds while i'm entering an answer to save tavern time - but there are times when I for instance get information for a particular column that must be right for two different reasons (the only non-eliminated gift in the specific column, and the only column where a gift in the wrong position could still go) and that raises the question of if my prior guess was optimal or if i could've used one of the columns to derive additional information.
Even on the second turn, after turn 1 say returns 4 wrong positions and 1 elimination, is it really best to try the remaining 5 options on turn 2 to find out if any of them are there (which is what I currently do) or is it perhaps better to only try 2 or 3 of the unknowns and try to find out some extra information on position for some of the ones you know are there but not where?
Another interesting question is if you're extremely short in a particular good how much does it hurt your success rate to just assume that good wasn't chosen. With 10 options and each option just being random[0,9], it's a 59% chance your assumption is correct - but some of the other 41% you might pick up anyways just by successfully eliminating all the alternatives and knowing exactly where to place your short good without having to try it in multiple spots. And a higher portion of the 59% than usual will be picked up because finding 5 from 9 options is an easier task than 5 from 10 in the same number of turns.
It would be very possible to build a program to answer these sorts of questions by brute force (reverse solving from "endgame" positions), and I expect some of them to be counterintuitive.
In our negotiation game, the answer to every query is independant - so there's no hidden biasing regarding things you did not ask about.
Some other tasks that might be of interest:
Now one thing that would be worth working on is ideal arrangements for partitioning the results with your 2nd last guess. I'm not 100% convinced that the obvious tries there are the best tries to optimise probability. I've done well enough so I haven't really approached the problem yet beyond what i could work out in my head in the 5 seconds while i'm entering an answer to save tavern time - but there are times when I for instance get information for a particular column that must be right for two different reasons (the only non-eliminated gift in the specific column, and the only column where a gift in the wrong position could still go) and that raises the question of if my prior guess was optimal or if i could've used one of the columns to derive additional information.
Even on the second turn, after turn 1 say returns 4 wrong positions and 1 elimination, is it really best to try the remaining 5 options on turn 2 to find out if any of them are there (which is what I currently do) or is it perhaps better to only try 2 or 3 of the unknowns and try to find out some extra information on position for some of the ones you know are there but not where?
Another interesting question is if you're extremely short in a particular good how much does it hurt your success rate to just assume that good wasn't chosen. With 10 options and each option just being random[0,9], it's a 59% chance your assumption is correct - but some of the other 41% you might pick up anyways just by successfully eliminating all the alternatives and knowing exactly where to place your short good without having to try it in multiple spots. And a higher portion of the 59% than usual will be picked up because finding 5 from 9 options is an easier task than 5 from 10 in the same number of turns.
It would be very possible to build a program to answer these sorts of questions by brute force (reverse solving from "endgame" positions), and I expect some of them to be counterintuitive.
DeletedUser31440
@xivarmy thanks for writing out that expanation for why Monty Hall doesn't apply, intuitively it makes sense that it wouldn't, but I'll be darned if I had the right words to explain why it doesn't work.
Now one thing that would be worth working on is ideal arrangements for partitioning the results with your 2nd last guess. I'm not 100% convinced that the obvious tries there are the best tries to optimise probability. I've done well enough so I haven't really approached the problem yet beyond what i could work out in my head in the 5 seconds while i'm entering an answer to save tavern time - but there are times when I for instance get information for a particular column that must be right for two different reasons (the only non-eliminated gift in the specific column, and the only column where a gift in the wrong position could still go) and that raises the question of if my prior guess was optimal or if i could've used one of the columns to derive additional information.
Even on the second turn, after turn 1 say returns 4 wrong positions and 1 elimination, is it really best to try the remaining 5 options on turn 2 to find out if any of them are there (which is what I currently do) or is it perhaps better to only try 2 or 3 of the unknowns and try to find out some extra information on position for some of the ones you know are there but not where?[/QUOTE]
I had been doing what you were talking about on the third turn. Not using a UUC or a PCUC item, but instead using a KC in a column if it only had Wrong Choices before. In that case I think the additional information you gain from using the KC is more valuable than the information you gain from eliminating a UUC or PCUC and having a column with no information on the KC's that aren't wanted for it. I started doing that a few weeks after writing this guide and I liked the results, but it wasn't a huge swing in getting more encounters done in 4 turns so I never added it to the guide.
This is interesting, I however don't think I'll be the guinnea pig on trying this out. For the program I tried building something in excel, but it ended up getting way more complicated then I had anticipated and I gave it up. I think I finished it out to the point of making guesses for the second round, the third round had too many dependancies and I was getting lost when writing my if/thens.
Now one thing that would be worth working on is ideal arrangements for partitioning the results with your 2nd last guess. I'm not 100% convinced that the obvious tries there are the best tries to optimise probability. I've done well enough so I haven't really approached the problem yet beyond what i could work out in my head in the 5 seconds while i'm entering an answer to save tavern time - but there are times when I for instance get information for a particular column that must be right for two different reasons (the only non-eliminated gift in the specific column, and the only column where a gift in the wrong position could still go) and that raises the question of if my prior guess was optimal or if i could've used one of the columns to derive additional information.
Even on the second turn, after turn 1 say returns 4 wrong positions and 1 elimination, is it really best to try the remaining 5 options on turn 2 to find out if any of them are there (which is what I currently do) or is it perhaps better to only try 2 or 3 of the unknowns and try to find out some extra information on position for some of the ones you know are there but not where?[/QUOTE]
I had been doing what you were talking about on the third turn. Not using a UUC or a PCUC item, but instead using a KC in a column if it only had Wrong Choices before. In that case I think the additional information you gain from using the KC is more valuable than the information you gain from eliminating a UUC or PCUC and having a column with no information on the KC's that aren't wanted for it. I started doing that a few weeks after writing this guide and I liked the results, but it wasn't a huge swing in getting more encounters done in 4 turns so I never added it to the guide.
This is interesting, I however don't think I'll be the guinnea pig on trying this out. For the program I tried building something in excel, but it ended up getting way more complicated then I had anticipated and I gave it up. I think I finished it out to the point of making guesses for the second round, the third round had too many dependancies and I was getting lost when writing my if/thens.
DeletedUser
I have actually done this frequently. It has a significantly better than 50% success rate in my experience, but it is extremely frustrating when you hold off on that short good only to find out it is the correct choice for one slot...or worse, multiple slots. Not to mention that when you're stubborn like I am, you end up spending Diamonds for that (those
) extra turn(s) when you finally give in and have to use that short good. DeletedUser35894
That’s the part I missed. That bit makes perfect sense now.
Why not edit the OP and take out the whole UUC vs. PCUC? It’s irrelevant, we have known and unknowns as has been pointed out and you seem to agree. It throws off much of the rest of the guide and adds confusion, especially for newbs or people not naturally inclined towards this sort of thing, which is who a guide like this would be for.
Last edited by a moderator:
DeletedUser31440
Even odds, so it doesn't hurt and it speeds up the decision process if you have a set system in place.
DeletedUser35894
The guide says they aren’t even so why leave incorrect information in that can only serve to confuse things?
People would choose cheapest goods, as you recommend, to speed up their decision process or some other method of their choosing.
You ask for people to point out the problems and make suggestions, but seem defensive about including them.
DeletedUser31440
There's a disclosure right below saying that UUC may not be less likely than PCUC, as it still is part of the system I use when negotiating I'm keeping it in. If it lessened odds I would take it out, it doesn't so I leave it in. If you want to use the cheapest option as your system feel free to do so, I find PCUC use to be faster.
DeletedUser35894
Yours is the most valuable info in the thread from what I know of statistics. I have a question that can help narrow down a simpler method.
So say you have a 6 choice encounter. You get 3 incorrect and two wrong.
You now have one UC. So you put it under one of the wrong choice slots. This leaves you with 4 slots and two KC’s.
3 of the slots are previously “Incorrect” and 1 is a “Wrong Person”.
So the two items that were wrong person choices now go under two of the incorrect slots as you mention. Is there any way to prioritize our choices at that point as we can’t avoid having 2 doubles at that point since our UUC is always supposed to go under the wrong person choice we put it there. Should we then also put it in one of the incorrect slots so we don’t have a double under an incorrect slot and let the other wrong person choice get doubled there?
I think a really simple guide could be done for this. I had found one here, but it was burried in another thread and can’t find it.
Basically we have KC and UC. A UC is by definition unused. Had it been used it would be either a Wrong Person or Incorrect. We’d have our answer.
All that’s needed is a set of rules and then which prioritize over the others.
Something like:
-KC’s go under Incorrect
-UC’s go under Wrong Person
-In your second through fourth turns avoid doubles where possible prioritizing not doing it under Incorrect over Wrong Choice (this is what I’m asking about and don’t know if it would be correct or not)
-Prioritize bringing out UUC’s even before laying down an item you know is correct as it’s more valuable to eliminate more items
-Recognize when you have all 5 KC’s come out. This wouldn’t have to be all on the first turn as the guide mentions?
-Work from lowest to highest probabilities
My only questions would be where to prioritize avoiding doubles and if it’s only more relevant to incorrect collums and possibly if certain actions trump others in all cases.
DeletedUser36713
I think that negotiations isn fair at all.
When negotiator answering with INCORECT or WRONG PERSON your goods are taken from you each time.
When negotiator answering with INCORECT or WRONG PERSON your goods are taken from you each time.
Emberguard
Well-Known Member
If they didn't take the goods, there'd be no cost and no risk involved.
When you fight with troops getting it "wrong" costs your troops lives. Why should negotiating be any different?
DeletedUser8428
I don't think it's unfair at all - your goods aren't "taken" from you - you used them in an attempt to negotiate your way through an encounter. Your attempt was not successful. It's like making a bet on a poker hand. If you don't win the hand, you lose the amount that you bet - that's how it works.
Penguin174
Member
Great reading but I still have an unanswered question. Does fighting or negotiating affect the points you earn per encounter or do you get the same points in either case??
DeletedUser29726
It does. They're supposed to be close, but it's not exactly the same. Also negotiating often involves giving up and restarting from time to time which is 'extra' points (you can do the same thing with fighting by killing some stuff and then surrendering/losing, but it's less typical in "natural" play than retrying a negotiation).
BruteForceAttack
Well-Known Member
negotiating gives more points then fighting, well rather negotiating all the way thru will end up more points, because when you fight even one hit on your army reduces your points.
DeletedUser37659
I hate to Negotiate, there needs to be a just kill them all button
Agent327
Well-Known Member
Thanks for that wonderfull contribution. Most useless necro I have seen in a while.
Grimmlion
New Member
GE is a must for me because I really enjoy it and the point is to have fun, right? GE also helps your guild. So I always have a ToR and level it up as a priority. Even if relics give you prizes you won't use, you can sell them for points and gems at auction. My highest city is at PMA so maybe it's different at higher levels but for my levels I think ToR is great.
When I started playing Youtube videos on GE negotiating strategy were a tremendous help.
When I started playing Youtube videos on GE negotiating strategy were a tremendous help.
WinnerGR
Well-Known Member
I take it your a trader to be in PMA and not fighting GE ?
Aurora Jolene
New Member
Here is my 2 cents....
1. In my opinion...(I don't know how to describe this so here is an example) say you have 2 spots correctly filled with coins, 1 spot correctly filled with supplies, and 1 spot correctly filled with medals. All supplies except coins and supplies are NOT an option. Then (imo) its more likely to be supplies. So if you have 2 correct of 1 of the same items and 1 correct with a different item and 1 spot left then it is more likely to be the one with 2 filled not 1. Hope that makes sense. Oh one more example could be showing the correct answers as A,B,C then it would be like A-C-B-A-B.
2. If you only have 3 options (or 4 if you use the extra turn boost) I like to first do all options as coins for example, because I have a huge amount of coins, then what's left I do all supplies because I have a lot of supplies (but less then coins) then I'll go on to the goods doing which ever one I have more of. It is sometimes wasteful but it definitely helps me use less of the good stuff...(there have even been a few times the answer was all coins so boo-ya!)
Also sidebar isn't it crazy when your on a really tough negotiation (like a high level 4 GE negotiation) and you get them all right on the first try! but the opposite is true and sucks when you have 1 slot left to fill and 2 options and you get it wrong
1. In my opinion...(I don't know how to describe this so here is an example) say you have 2 spots correctly filled with coins, 1 spot correctly filled with supplies, and 1 spot correctly filled with medals. All supplies except coins and supplies are NOT an option. Then (imo) its more likely to be supplies. So if you have 2 correct of 1 of the same items and 1 correct with a different item and 1 spot left then it is more likely to be the one with 2 filled not 1. Hope that makes sense. Oh one more example could be showing the correct answers as A,B,C then it would be like A-C-B-A-B.
2. If you only have 3 options (or 4 if you use the extra turn boost) I like to first do all options as coins for example, because I have a huge amount of coins, then what's left I do all supplies because I have a lot of supplies (but less then coins) then I'll go on to the goods doing which ever one I have more of. It is sometimes wasteful but it definitely helps me use less of the good stuff...(there have even been a few times the answer was all coins so boo-ya!)
Also sidebar isn't it crazy when your on a really tough negotiation (like a high level 4 GE negotiation) and you get them all right on the first try!
but the opposite is true and sucks when you have 1 slot left to fill and 2 options and you get it wrong