• 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:
    Apply

[Guide] Negotiating GE Tips

DeletedUser31440

I tend to negotiate in GE from level 3 to 4, sometimes the last half of level 2 depending on if I have enough time to fight through or just want to get through it quickly. Regardless, at some point almost everybody who pushes into the higher levels of GE will do some negotiating, so why not have a thread that gives general guidelines and tips on how to succesfully negotiate (minimize goods losses, minimize diamond expenditure, and decrease time spent negotiating). I'm going to compile several tips and my thought processes for why I use them, along with pictures to hopefully clarify what I'm talking about.

It should also be known that some encounters will be impossible to guarantee a correct combination in 4 turns, the best that can be hoped for is luck and minimizing the damage in those cases.

There are a few terms I'll be using throughout, I'll try to expain them below:

Choice - goods/supplies/coins/medals - what you pay during the encounters

Known Choice (KC) - A choice that you know will be part of the correct answer in an encounter. These are the choices that gave you a wrong person answer in previous guesses. In the below picture copper, honey and supplies are all KC's.
2.PNG

Unused Unknown Choice (UUC) - These are choices that you have not yet used and are unsure if they will be a part of the correct answer. In the below picture coins and supplies are both UUC's.2a.PNG

Previously Correct Unknown Choice (PCUC): These are choices that only have a correct answer from previous attempts. They could still be used again, but they have been correct in previous attempts. In the picture above honey and glass are both PCUC's.

General Guidelines:

If you know that you will be negotiating more than 4 turns activate the Extra Turn boost from the tavern.
I personally use this whenever negotiating, regardless of the level of GE that I'm currently on. I do this to minimize goods losses and to ensure that even if I fat finger something and mess up a round I will generally still be able to complete the encounter without using diamonds.​

If you are not on your last turn and you have unknown choices remaining it is better to eliminate possible answers than get correct answers.
While it may be more immediately rewarding to get a correct answer that will oftentimes come back to haunt you later in the negotiation. In the below picture I know that salt has to be in one of the two open spots, but I still have 2 PCUC's honey and supplies. To figure out which of the other 2, if either, is in the final answer I will not use salt as part of my answer. (Yes, I realize this isn't the best illustration of this tip, I have enough turns left to use salt in the first guess and still guarantee a correct answer, but for illustration purposes it works)​
3b.PNG

A UUC is more likely to be a correct choice than a PCUC, if you have 3 open slots, 3 UUC's and 2 PCUC's, try the 3 UUC's before using the PCUC's.
Full disclosure, I'm not sure that the math always agrees with me on this one, but as I value speed of negotiations I use this as a general rule instead of figuring out percentage chances each time when doing a negotiation.​

Always try to place an PCUC and/or UUC under a wrong person. PCUC's get priority over UUC's when going under a wrong person.
You should try to minimize the number of wrong answers in a given column. A PCUC should be more likely to give you an incorrect answer than a UC so they should be placed under the wrong person instead of placing a UC under the wrong person.
Recognize when you have all of the available choices as KC's.
There will be times when your first guess yields you with 5 KC's. If the negotiation has more than 5 choices available it is no longer necessary to eliminate the other choices. You known all 5 KC's so only use those 5 KC's to solve the rest of the negotiation, using a UUC in this case will be wasting turns.​

Work from the highest probability slot to the lowest.
Oftentimes you will have multiple slots still open on the 4th turn of a negotiation. When this is the case, or whenever you have all remaining KC's, work from the most likely correct answer to the least likely. In the below picture the far right column has a 100% chance of being gold. That should be the first answer that you give. All the remaining columns have a 1 in 3 chance of being any KC. Bricks and medals have a 50 - 50 shot of going into 2 of the columns, those should be the next two choices that you make, copper has the lowest odds and should be the last KC placed.​
abc.JPG

If there are only 4 choices in an encounter use the least expensive choice in all slots for the first attempt, the second least expensive choice in all remaining slots for the second attempt, the third least expensive choice in all remaining slots for the third attempt and then the most expensive choice for any remaining slots on the 4th attempt.
This works to minimize losses and also dummy proofs the negotiation. The only possible way to get one of these encounters wrong is if you used all 4 choices in a round and then used a KC in the same column where it was a wrong person. If you only use one choice at a time across the board you will be unable to use a choice in the next attempt as it will either be a solved encounter or that choice will have rendered an incorrect response and won't be eligible for use again.​

Only use diamonds for an extra turn if you are 100% certain that you know the correct answer.
This rule is not hard and fast for everybody, some players may value their goods more than their diamonds, if that's the case completely disregard this. For me personally, I value diamonds higher than goods, so if I am not certain that the next answer will be correct I refuse to spend diamonds for another chance.​

That is all I have for now, if you see any fallacies in my logic, please let me know. If you have a tip that isn't listed, please let me know. Hopefully this is helpful for some people.
 
Last edited by a moderator:

DeletedUser31498

I tend to negotiate in GE from level 3 to 4, sometimes the last half of level 2 depending on if I have enough time to fight through or just want to get through it quickly. Regardless, at some point almost everybody who pushes into the higher levels of GE will do some negotiating, so why not have a thread that gives general guidelines and tips on how to succesfully negotiate (minimize goods losses, minimize diamond expenditure, and decrease time spent negotiating). I'm going to compile several tips and my thought processes for why I use them, along with pictures to hopefully clarify what I'm talking about.

It should also be known that some encounters will be impossible to guarantee a correct combination in 4 turns, the best that can be hoped for is luck and minimizing the damage in those cases.

There are a few terms I'll be using throughout, I'll try to expain them below:

Choice - goods/supplies/coins/medals - what you pay during the encounters

Known Choice (KC) - A choice that you know will be part of the correct answer in an encounter. These are the choices that gave you a wrong person answer in previous guesses. In the below picture copper, honey and supplies are all KC's.
View attachment 9193

Unused Unknown Choice (UUC) - These are choices that you have not yet used and are unsure if they will be a part of the correct answer. In the below picture coins and supplies are both UUC's.View attachment 9194

Previously Correct Unknown Choice (PCUC): These are choices that only have a correct answer from previous attempts. They could still be used again, but they have been correct in previous attempts. In the picture above honey and glass are both PCUC's.

General Guidelines:

If you know that you will be negotiating more than 4 turns activate the Extra Turn boost from the tavern.
I personally use this whenever negotiating, regardless of the level of GE that I'm currently on. I do this to minimize goods losses and to ensure that even if I fat finger something and mess up a round I will generally still be able to complete the encounter without using diamonds.​

If you are not on your last turn and you have unknown choices remaining it is better to eliminate possible answers than get correct answers.
While it may be more immediately rewarding to get a correct answer that will oftentimes come back to haunt you later in the negotiation. In the below picture I know that salt has to be in one of the two open spots, but I still have 2 PCUC's honey and supplies. To figure out which of the other 2, if either, is in the final answer I will not use salt as part of my answer. (Yes, I realize this isn't the best illustration of this tip, I have enough turns left to use salt in the first guess and still guarantee a correct answer, but for illustration purposes it works)​
View attachment 9195

A UUC is more likely to be a correct choice than a PCUC, if you have 3 open slots, 3 UUC's and 2 PCUC's, try the 3 UUC's before using the PCUC's.
Full disclosure, I'm not sure that the math always agrees with me on this one, but as I value speed of negotiations I use this as a general rule instead of figuring out percentage chances each time when doing a negotiation.​

Always try to place an PCUC and/or UUC under a wrong person. PCUC's get priority over UUC's when going under a wrong person.
You should try to minimize the number of wrong answers in a given column. A PCUC should be more likely to give you an incorrect answer than a UC so they should be placed under the wrong person instead of placing a UC under the wrong person.
Recognize when you have all of the available choices as KC's.
There will be times when your first guess yields you with 5 KC's. If the negotiation has more than 5 choices available it is no longer necessary to eliminate the other choices. You known all 5 KC's so only use those 5 KC's to solve the rest of the negotiation, using a UUC in this case will be wasting turns.​

Work from the highest probability slot to the lowest.
Oftentimes you will have multiple slots still open on the 4th turn of a negotiation. When this is the case, or whenever you have all remaining KC's, work from the most likely correct answer to the least likely. In the below picture the far left column has a 100% chance of being gold. That should be the first answer that you give. All the remaining columns have a 1 in 3 chance of being any KC. Bricks and medals have a 50 - 50 shot of going into 2 of the columns, those should be the next two choices that you make, copper has the lowest odds and should be the last KC placed.​
View attachment 9198

If there are only 4 choices in an encounter use the least expensive choice in all slots for the first attempt, the second least expensive choice in all remaining slots for the second attempt, the third least expensive choice in all remaining slots for the third attempt and then the most expensive choice for any remaining slots on the 4th attempt.
This works to minimize losses and also dummy proofs the negotiation. The only possible way to get one of these encounters wrong is if you used all 4 choices in a round and then used a KC in the same column where it was a wrong person. If you only use one choice at a time across the board you will be unable to use a choice in the next attempt as it will either be a solved encounter or that choice will have rendered an incorrect response and won't be eligible for use again.​

Only use diamonds for an extra turn if you are 100% certain that you know the correct answer.
This rule is not hard and fast for everybody, some players may value their goods more than their diamonds, if that's the case completely disregard this. For me personally, I value diamonds higher than goods, so if I am not certain that the next answer will be correct I refuse to spend diamonds for another chance.​

That is all I have for now, if you see any fallacies in my logic, please let me know. If you have a tip that isn't listed, please let me know. Hopefully this is helpful for some people.

Great idea to write this up, but this really stood out to me:
"A UUC is more likely to be a correct choice than a PCUC, if you have 3 open slots, 3 UUC's and 2 PCUC's, try the 3 UUC's before using the PCUC's."
"PCUC's get priority over UUC's when going under a wrong person. A PCUC should be more likely to give you an incorrect answer than a UC so they should be placed under the wrong person instead of placing a UC under the wrong person."

Why do you think this is true? You're saying that given gold is correct for someone, gold is less likely to be correct for someone else than a random other good? I think I've seen enough cases where there were 4 of the same good that this doesn't hold, but very curious if you have numbers on this being correct that all choices aren't independent of others.

also small correctoin but you mean far right" In the below picture the far left column has a 100% "
 

DeletedUser31440

@gutmeister I haven't done any statistics on it, but it tastes good in my mouth, if that makes sense. I also tend to run through negotiations as quickly as possible so I try to set definitive if then for my choices to speed up my decision making. Truthfully it should be looked at on a turn by turn basis, but for sake of time I use it as a rule of thumb.

Also I do tend to mix up my left and right an embarrassingly amount of times.
 

Woody*

Active Member
Great post! A couple comments..

A UUC is more likely to be a correct choice than a PCUC

Their probabilities are exactly the same.

Other tips: 1) put KCs in incorrect columns, and 2) in Rounds 2-4 avoid duplicating your guesses in multiple columns if you can avoid it. For example, consider a 4-option encounter where you have 1 Correct (a good), 2 Wrong Person (a different good and coins) and two Incorrect (Supplies). You still can guess either good and coins. For the 2 "incorrect" supply guesses in Round 1, avoid guessing coins on both (or one of the goods for both).
 

DeletedUser35894

I'm probably just misunderstanding you, but in your second example, why are supplies and coins UUC's? In the first example, you state, and I understand that a wrong person gives you a KC, but here it's considered UUC?...and coins got incorrect and there still called unused and unknown? They seem used to me and once incorrect they are off the table for any slot.

So again, I'm sure I'm just misunderstanding, but after that point it makes the guide hard to follow when I'm unclear what you're saying.

There was another guide and now I can't find it, but was pretty simple to follow and included options for what to do as more goods/supplies come into play.

Thanks for the work on this. People put a lot of time into these things and don't get much thanks. Also, surprising how few folks in the game bother to learn much about when they're spending much time on it and often a lot of money.
 

DeletedUser35894

Examples are stand alone, the picture for KC has no relevance to the UUC example.

Thanks, but then I'm still lost because my reply didn't assume such as they have the same turn amounts left, I treated them as independent. I don't understand how the same answer of "wrong person" can give both the outcome of KC and UUC. As for the coins, they are definitely used, they are out of the equation as incorrect. The glass isn't any less likely to be correct because of it's status as having been correct on another slot.
 

DeletedUser31440

Thanks, but then I'm still lost because my reply didn't assume such as they have the same turn amounts left, I treated them as independent. I don't understand how the same answer of "wrong person" can give both the outcome of KC and UUC. As for the coins, they are definitely used, they are out of the equation as incorrect. The glass isn't any less likely to be correct because of it's status as having been correct on another slot.

UUC example, coins and supplies haven't been guessed yet, only chosen but not yet submitted to tribal folk.
 

Dido 815 the Scourge

Active Member
Just wanted to say that this guide has helped me TREMENDOUSLY! My city & buffs are in no way ready to fight through all of level 3, much less even trying to fight in level 4. Thanks to your guide, I’ve been able to complete all 4 GE maps with a reasonable use/loss of goods/supplies/coins/medals. This has also helped my level 6 ToR payoff wonderfully (4X 100 FP gold relics in the past 2 week’s, tyvm).

One thing I did notice and not sure if its purely anecdotal or what: when I get to the last turn and I have to pick between 2 goods (goods only, nothing else), I always guess correctly when I use the good that I have the most of between the two.

E.g, I’ve got 60 basalt and 55 paper. Either could be correct. I guess the higher of the two (basalt), and I’m correct.

Honestly can’t think of a time when this HASN'T worked. Doesn’t happen often and just wonder if anyone has had the same experience.

Again, TYVM for this guide!
 

DeletedUser31440

@Dido 815 the Scourge, I'm glad that you've found this helpful. As for the always being the higher of 2 unknown goods I can say quite definitely that is not always the case, but it is nice when it turns out to be.
 

DeletedUser31498

@Claire the Divergent and others with strong statistical knowledge: this is the one point I'm not sure of. Let's say in Round 1 I get coins as a wrong person. Round 2 I place coins correctly.

Is this a Bayesian situation where given that coins was "wrong" before, and "Correct" in round 2, presented with the situation where I am deciding between coins and other unknown items in Round 3 (like a previous resource which was "correct" on first try), is it more likely to be a correct answer than the others? I think we all agree that the UUC is equally likely as a PCUC going forward, but is there information in the wrong answer even after it's revealed as correct??

Please, if you're not very familiar with statistics and Bayes in particular don't weigh in on this particular question. thanks!!

@ThruTHEhead any thoughts on this?
 

DeletedUser31440

@gutmeister is this the situation you're talking about?

1.JPG

Edit: Looking at it again I don't think it is, it would need another option that hasn't been used yet to fit.
 
Last edited by a moderator:

DeletedUser31440

@gutmeister I think this is what you're talking about. It's the option between marble and limestone for column C and D in the below.

Negotiation 1.PNG
Negotiation 2.PNG
Negotiation 3.PNG

If that's the question, there's no pre-existing information known that would effect either option. If I were actually negotiating it out I'd do limestone in A, marble in C, and ebony in D; also would have the 4th turn active too though.
 

Woody*

Active Member
Disclaimer...it's been too long since I've studied Bayesian statistics. But if we assume all choices have equal probability and all choices are truly random, I don't see how there would be any difference in probability.
 

DeletedUser31498

@gutmeister I think this is what you're talking about. It's the option between marble and limestone for column C and D in the below.

View attachment 10727
View attachment 10728
View attachment 10729

If that's the question, there's no pre-existing information known that would effect either option. If I were actually negotiating it out I'd do limestone in A, marble in C, and ebony in D; also would have the 4th turn active too though.

Yes very similar to that situation:
In mine, are coins or paper more likely to be in Column C. THe information we have is Coin was previously wrong, which we later found a home for while Paper was simply Correct. Is there any information?
 

Attachments

  • GE_neg.png
    GE_neg.png
    455.7 KB · Views: 17

Woody*

Active Member
Yes very similar to that situation:
In mine, are coins or paper more likely to be in Column C. THe information we have is Coin was previously wrong, which we later found a home for while Paper was simply Correct. Is there any information?

There is no information on what Column C will require. Coin and Paper are equally likely (with the assumptions I make above, and the added assumption that earlier guesses don't influence future unrelated guesses...and frankly it would take work on the programmers' part to put non-randomness in...which I can only imagine they would consider doing to influence diamond-related events).
 

DeletedUser31440

Yes very similar to that situation:
In mine, are coins or paper more likely to be in Column C. THe information we have is Coin was previously wrong, which we later found a home for while Paper was simply Correct. Is there any information?

I agree with @Claire the Divergent on this one, there's no pertinent information to effect your choice in column C. This is the breakdown of how I see it, if someone has a better way of looking at it, please let me know.

Column C - 50/50 between Paper and Coins
Column D - 33/33/33 between Paper, Coins and Basalt

Basalt - 100% chance included in answer, 100% chance of being in Column D
Paper - 50% chance included in answer, 100% chance of being in Column C if in answer
Coins - 50% chance included in answer, 100% chance of being in Column C if in answer

Actual guess if I was negotiating it:
Column C - Coins (cheaper than paper and same odds)
Column D - Basalt
 

DeletedUser31498

Right I get intuitively it seems like there isn't information. The Monty Hall problem went un-noticed for literally years, even though it's amazingly obvious in hindsight. So I'm really just reaching out to those with strong probability backgrounds if there's something subtle here. Not that you all are wrong, just that the obvious answer can definitely be wrong. But I do agree with you all, I think this is not a Bayesian situation.
 

DeletedUser29726

Well the obvious answer is just to write out every answer which still remains possible - which without any additional information on generation conditions should all have been equally likely to begin with, and must all be equally likely now as they have not been eliminated. You then would then count the number of occurrences of column C being each item to determine the probability of what it is.

However, it's not impossible generation conditions are biased to prefer first occurrences over repeat occurrences - i don't think i've ever seen a 5 of the same be the correct solution even though under unbiased conditions that should happen 1/32 times in encounter 1 and 2. There's not really a good way to test how they generate the solutions though - so assuming it's unbiased is probably the best we can do. But if you want to believe an item that was correct is less likely than item you've just never tried, you're free to do so - in the expected scenario where neither is preferred you're doing no damage. If they prefer repeats over first occurrences in generation (seems unlikely), then you'd be hurting yourself.
 
Top