Empty Slot Merchant Placement Success

[ws2_32]

According to interpretation of the 1.09 release notes, the latest formula for empty slot merchant placement success should be:

Version 1.09 empty slot merchant placement success% =
35% +
(5% * stability) +
(3% * ADM) +
(5% if the CoT is owned by merchant sender) +
(50% * trade efficiency)

Minimum empty slot merchant placement success = 25%
Maximum empty slot merchant placement success = 95%

The difference between 1.09 and 1.08 versions is mainly that trade efficiency (TE) helps determine how successful merchant placement will be. For some nations with low TE, merchant placement will be more difficult than in 1.08. For example, a nation with TE of 10% in 1.09 will have an empty slot merchant placement success rate between 25% and 87% whereas in 1.08 the range was 35% to 95% regardless of any TE below 100%. A nation with TE of 50% in 1.09 (close to Venice, Genoa, and Portugal's GC TE level) will have an empty slot success rate between 45% and 95%.

Game play with regards to merchant placement should not be dramatically different than in in versions prior to 1.09. Overall, expect an improvement in merchant placement success. The degree of improvement in merchant placement success now depends much more on small changes of TE, since TE of 100% is no longer required to cause a change in success rates.


Testing was only done on the 1.08 vanilla. The tests showed the previous formula was:

Version 1.08 empty slot merchant placement success% =
50% +
(5% * stability) +
(3% * ADM) +
(5% if the CoT is owned by merchant sender) +
{50% * truncate (trade efficiency)}

Minimum empty slot merchant placement success = 35%
Maximum empty slot merchant placement success = 95%

If stability is negative, then stability has a negative effect on the chance of empty slot merchant placement success.

The monarch's current effective administrative rating, ADM, is in the range of 0 to 9.

There is a 5% improvement in empty slot merchant placement success if the CoT where the merchant is placed is owned by the country making the placement.

There was an apparent bug in 1.08 where trade efficiency was truncated to an integer within the computation. This type of error is known as a typecast error. The end result is that, in regards to empty slot merchant placement success, there was no difference between 20% and 99% trade efficiencies. Similarly, there was no difference between 100% and 199% trade efficiencies. There were however great differences between 99% and 100% trade efficiencies and between 199% and 200% trade efficiencies. By reaching 200% trade efficiency, empty slot merchant placement success was the maximum of 95% success rate no matter what the other factors.

a) truncate (20%) = truncate (0.20) = 0
truncate (99%) = truncate (0.99) = 0

b) truncate (100%) = truncate (1.00) = 1
truncate (199%) = truncate (1.99) = 1

c) truncate (200%) = truncate (2.00) = 2
truncate (299%) = truncate (2.99) = 2

The truncated numbers were returned before multiplying by 50%, which is why there was such a big difference between 0.99 trade efficiency and 1.00 trade efficiency.

No other known factors influence empty slot merchant placement success. Tested are: Mercantilism DP slider, trade tech level, badboy points, trade tech level (different than trade efficiency but still influencing trade efficiency), slot number to be filled, distance between CoT and capital, continent of capital, whether or not merchant sender has a province that trades in the destination CoT, religious difference between merchant sender and CoT owner, war between merchant sender and CoT owner, and relations between merchant sender and CoT owner.

[RichardTheFirst]

What influences trade efficiency? Trade level seems to have just some influence...

Useful stuff. Congrats!

[DSYoungEsq]

Tom, I have just one question about this study:

What happens when more than one merchant is sent to a CoT with only one placement slot available, and both countries are "successful" in placing the merchant?

[the_genius]

What happens when more than one merchant is sent to a CoT with only one placement slot available, and both countries are "successful" in placing the merchant?

Shouldn't be the times when a merchant kicks another out?

Btw, nice study Tom...

That issue on the TE is pretty bad, since a lot of people relies on that to access succesfully into a CoT... also, about the DP Trade slider, doesn't has an influence in the recent betas?

[Miozozny]

There are however great differences between 99% and 100% trade efficiencies and between 199% and 200% trade efficiencies. By reaching 200% trade efficiency, empty slot merchant placement success is the maximum of 95% success rate no matter what the other factors.

If I understand it all correctly, the difference between 199% and 200% isn't nearly as big as between 99% and 100%. At 100% the placement chance is at least 85% (88% under normal circumstances (ADM > 0)). So going to 200% would only increase this chance by at most 10%.

Is my conclusion correct or am I missing something?

BTW; very useful study. 100% trade efficiency has suddenly become an important goal. To me anyway.

[ws2_32]

What influences trade efficiency? Trade level seems to have just some influence...

Trade tech level, Aristocracy DP slider, Land Dp slider, and refineries all influence trade efficiency.

What happens when more than one merchant is sent to a CoT with only one placement slot available, and both countries are "successful" in placing the merchant?

That is beyond the scope of this study. To automate this study I created a "sandbox" of 20 empty European CoTs with a limited number of countries. None of the other countries were able to trade at all. The other four countries were held at trade tech zero by setting the date for reaching trade tech one to 9999. I did 27,600 merchant placements with autosend. At the end of it all, I noticed that some placements had to be excluded because the autosend sometimes placed merchants in CoTs that already reached monopoly status. When that was taken into consideration, the numbers fitted the nice neat formula much better.

Shouldn't be the times when a merchant kicks another out?

Btw, nice study Tom...

That issue on the TE is pretty bad, since a lot of people relies on that to access succesfully into a CoT... also, about the DP Trade slider, doesn't has an influence in the recent betas?

I have not checked the kicking out of other merchants yet. Yes, the TE bug is really a disappointment now that there are not going to be new patches. I suspect that if it were to be corrected the base 50% success rate would need to be a little lower, like 25% to 35%.

I assume you are talking about the Mercantilism DP slider in the recent betas. My earlier testing on the trade systems between 1.08 and 1.08 beta has shown little difference. But I could run some tests someday. Regardless, I suspect little difference in Mercantilism DP slider since the effect of the slider has only been lessened in 1.08. There is no knowing yet if Mercantilism helps kick merchants out of CoTs.

If I understand it all correctly, the difference between 199% and 200% isn't nearly as big as between 99% and 100%. At 100% the placement chance is at least 85% (88% under normal circumstances (ADM > 0)). So going to 200% would only increase this chance by at most 10%.

Is my conclusion correct or am I missing something?

BTW; very useful study. 100% trade efficiency has suddenly become an important goal. To me anyway.

Yes, you have the basics of it. But understand that even below 100% trade efficiency, at high stability, and with a decent ADM, you have a placement success in the range of 75% to 85%. While 100% trade efficiency surely is nice to help you in times of low stability, really it is not generally of such great importance. However, if you are anywhere near 100% trade efficiency, I suggest doing what you can to get the few more percentage points of trade efficiency as soon as possible.

[RichardTheFirst]

If I understand it correctly you guys are disapointed because the chances of placement success are the same for 20% or 99% TE.

But what about chances of kick-out of the CoT? I think the reason why you stop investing in merchant placement when you have a low TE, is not the chances of placement of a merchant. Is the chances that merchant will be kicked-out too soon, not returning the investment.

A study about chances of kick-out would be interesting, don't you think?

[Ironfoundersson]

Intersting stuff

[ws2_32]

If I understand it correctly you guys are disapointed because the chances of placement success are the same for 20% or 99% TE.

But what about chances of kick-out of the CoT? I think the reason why you stop investing in merchant placement when you have a low TE, is not the chances of placement of a merchant. Is the chances that merchant will be kicked-out too soon, not returning the investment.

A study about chances of kick-out would be interesting, don't you think?

I already did some preliminary testing on "merchant stickiness" which I consider different than "merchant kick-out". The preferred method of placing merchants is through empty slot placement. Next, you want the merchants to stick. Luckily, trade efficiency affects the merchant’s stickiness in a way that it should. Mercantilism has the opposite effect as documented in release notes. Each trade level has an effect on merchant stickiness equal to about 3% of trade efficiency that is in addition to the trade efficiency a tech level already adds. There is a link in my signature.

Kicking merchants out of CoTs requires a different study. I am not sure how valuable a study on that is considering it is not a good way to get merchants into CoTs. However, we could learn something that could allow determining when there might sometimes be easy targets for expulsion from CoTs.

[RichardTheFirst]

Kicking merchants out of CoTs requires a different study. I am not sure how valuable a study on that is considering it is not a good way to get merchants into CoTs. However, we could learn something that could allow determining when there might sometimes be easy targets for expulsion from CoTs.

Yes. That study would really complete the issue here. Go for it!

CoT competiveness must also have a major role in here besides own TE and other's TE, IMHO. I don't know how we can determine which CoT's are more competitive because I suppose it all depends on the number of countries accessing the particular CoT, if they can access others, the number of merchants they have, etc, etc, etc.

Give you an example, as Portugal in middle 1500's I own Zacatecas, but my merchants don't stick in there more than some monthes. My TE is about 31% and other competitors (Maya, Lenape, Navajo, Cherokee, etc.) vary from 20 to 30%. I suppose the reasons are:
- it's the most valuable CoT in the world, the other two CoT's in the Americas are the less valuables in the world.
- they don't "see" many other CoT's so they are sending their merchants always to the same one
- they get a lot of merchants per year.

[ws2_32]

RichardTheFirst,

You could really benefit from skimming through my merchant stickiness studies. In the "Important Observations" section, I describe that a pagan nation (Xhosa) with 20% trade efficiency had a far superior advantage in sticking in Zacatecas when compared with Europeans with over 50% trade efficiency. Only two or three of the American nations that knew of the CoT in Zacatecas had begun trading since it takes some time for them to reach trade tech one. Still, one or two of the Europeans were sometimes completely expelled. Xhosa always stood strong in Zacatecas with four or five slots filled. I suspect religion has something to do with competition in CoTs.

Competition is a much harder thing to dissect. I do not know how the mechanism works. Is there first a match-up of one trader to another and then a roll to see who stays or goes? Or is there first a role to see if a merchant will be successful, and then if successful a separate roll to see who will go?

[broadsword]

Empty slot merchant placement success% =
50% +
(5% * stability) +
(3% * ADM) +
(5% if the CoT is owned by merchant sender) +
{50% * truncate (trade efficiency)}
.
.
.
There is an apparent bug where trade efficiency is truncated to an integer within the computation. This type of error is known as a typecast error. The end result is that, in regards to empty slot merchant placement success, there is no difference between 20% and 99% trade efficiencies. Similarly, there is no difference between 100% and 199% trade efficiencies. There are however great differences between 99% and 100% trade efficiencies and between 199% and 200% trade efficiencies. By reaching 200% trade efficiency, empty slot merchant placement success is the maximum of 95% success rate no matter what the other factors.

I don't know what you mean by truncated (trade efficiency).

Please indicate the truncated trade efficiencies of the following:
a) TE 1-99%,
b) TE 100-199%,
c) TE 200-299%.

For clarity, are the numbers that you return before or after multiplying by 50%?

Was this testing done on the latest beta?

[dsk]

Empty slot merchant placement success% =
50% +
(5% * stability) +
(3% * ADM) +
(5% if the CoT is owned by merchant sender) +
{50% * truncate (trade efficiency)}

Minimum empty slot merchant placement success = 35%
Maximum empty slot merchant placement success = 95%

If stability is negative, then stability has a negative effect on the chance of empty slot merchant placement success.

The monarch's current effective administrative rating, ADM, is in the range of 0 to 9.

There is a 5% improvement in empty slot merchant placement success if the CoT where the merchant is placed is owned by the country making the placement.

There is an apparent bug where trade efficiency is truncated to an integer within the computation. This type of error is known as a typecast error. The end result is that, in regards to empty slot merchant placement success, there is no difference between 20% and 99% trade efficiencies. Similarly, there is no difference between 100% and 199% trade efficiencies. There are however great differences between 99% and 100% trade efficiencies and between 199% and 200% trade efficiencies. By reaching 200% trade efficiency, empty slot merchant placement success is the maximum of 95% success rate no matter what the other factors.

No other known factors influence empty slot merchant placement success. Tested are: Mercantilism DP slider, badboy points, trade tech level (different than trade efficiency but still influencing trade efficiency), slot number to be filled, distance between CoT and capital, continent of capital, whether or not merchant sender has a province that trades in the destination CoT, religious difference between merchant sender and CoT owner, war between merchant sender and CoT owner, and relations between merchant sender and CoT owner.

Edit: Trade tech level has no effect either.

How did you work out that it was a typecast error? By educated guess or decompiling the binary? :curious:

[RichardTheFirst]

I suspect religion has something to do with competition in CoTs.

Well, if that's the case, it does not make any sense. Pagan nations were by far more underdevoloped in trade terms in this period. And big colonizers dominated the trade in and from all regions they colonised without the need to completely conquer these nations to achieve that objective. In my case Portugal have more than 60% of the provinces of America, 30% belong to pagan nations and Spain and 10% are still not colonized. However Portugal have only less than 5% of the Trade! Doesn't make sense and its a big minus point for the game IMO.

Competition is a much harder thing to dissect. I do not know how the mechanism works. Is there first a match-up of one trader to another and then a roll to see who stays or goes? Or is there first a role to see if a merchant will be successful, and then if successful a separate roll to see who will go?

Yes, I agree it will be very dificult to reach any conclusions on this. The variables and the different situations that could arise are just too many.

[ws2_32]

RichardTheFirst,

I think the best I can do is rate things in terms of one common factor, that being trade efficiency. So for example, one level of trade tech difference counts for 3% of trade efficiency in terms of competitiveness. Two movements of the Mercantilism DP slider may be worth about the same 3% in trade efficiency. But I may find those factors to have more or less impact depending on the relative trade strengths of other nations. Also, how about the other factors or suspected factors: CoT ownership, ADM, badboy, stability, religion, slot number, etc.?

I don't know what you mean by truncated (trade efficiency).

Please indicate the truncated trade efficiencies of the following:
a) TE 1-99%,
b) TE 100-199%,
c) TE 200-299%.

For clarity, are the numbers that you return before or after multiplying by 50%?

Was this testing done on the latest beta?

a) truncate (1%) = truncate (0.01) = 0
truncate (99%) = truncate (0.99) = 0

b) truncate (100%) = truncate (1.00) = 1
truncate (199%) = truncate (1.99) = 1

c) truncate (200%) = truncate (2.00) = 2
truncate (299%) = truncate (2.99) = 2

The numbers are returned before multiplying by 50%, which is why there is such a big difference between 0.99 trade efficiency and 1.00 trade efficiency.

Testing was only done on the 1.08 vanilla.

How did you work out that it was a typecast error? By educated guess or decompiling the binary? :curious:

That is an educated guess. Obviously the programmer meant for trade efficiency to influence the success rate of placing merchants. But it does not make sense that it should be so drastic at only 1.00 and 2.00 trade efficiencies. So certainly there is a casting where trade efficiency gets changed to an integer, which is why the effect is at the boundaries of 1.00 and 2.00. And because C++ is intentionally so loosely typed, these errors go unreported during compilation.

And when I say that the base 50% should be changed if the typecast error is corrected, I mean that the initial 50% should be lowered to 25% to 35% not the 50% value which is multiplied by the trade efficiency. If the base 50% was not lowered, you can see that right off the bat, success rates would be far greater than in current game play.

[RichardTheFirst]

I think the best I can do is rate things in terms of one common factor, that being trade efficiency. So for example, one level of trade tech difference counts for 3% of trade efficiency in terms of competitiveness. Two movements of the Mercantilism DP slider may be worth about the same 3% in trade efficiency. But I may find those factors to have more or less impact depending on the relative trade strengths of other nations. Also, how about the other factors or suspected factors: CoT ownership, ADM, badboy, stability, religion, slot number, etc.?

Hmmm I suspect that analysis will be very incomplete. Even with a lot of advantage in terms of TE or TL we saw that in a very competitive CoT one have very few sticking chances.

What if you study stickness just in very stable environments? For example Ivoria or Zimbabwe (or Zacatecas or Manhattan before colones arrive). You can then reach conclusions without the competitiveness factor.

Competiveness will then be an unexplained we know that exists we know we can't survive avoid this CoT factor. Players will know the general rules of the game except for the cuthroat environments.

EDIT - Oh, one more thing: I know from experience stability and ADM have influence in stickness. Don't know for sure about the other factors you mentioned but I suspect religion is not a factor here.

[DSYoungEsq]

EDIT - Oh, one more thing: I know from experience stability and ADM have influence in stickness. Don't know for sure about the other factors you mentioned but I suspect religion is not a factor here.

A mild correction I will offer here. You do not KNOW anything. The plural of anecdote is not evidence.

However, many of us strongly suspect that stability, at least, affects staying power and ability to get back in the door. We simply see all too often the result of an unfortunate Political Crisis.

[RichardTheFirst]

A mild correction I will offer here. You do not KNOW anything. The plural of anecdote is not evidence.

However, many of us strongly suspect that stability, at least, affects staying power and ability to get back in the door. We simply see all too often the result of an unfortunate Political Crisis.

Man, you must have been reading Plato a lot lately ! He's the one who said "all I know is I know nothing". I'm not that fanatical about this phrase of him although I see its merits. So I repeat "I know from experience" and I keep what I said.

What is knowledgement but the acumulation of experience? Statistics just help us with more quantitative data, but they are not knowledgement by itself, they are a support to knowledgement. If I eat one chicken per week and you eat none that doesn't mean we both eat half a chicken per week .

If we follow you or Plato we would never "know" nothing.

Be well, my friend .

[ws2_32]

I think therefore I am. That is all I know. But really how do you explain the staying power of pagans in distant pagan CoTs while other better traders are consistently being kicked out? If it is not religion then there is something else at work. From my examination of my Oman trials, I did not see staying power of Oman in Alexandria where Mameluks has the same Shiite religion. So I am baffled for now.

Anyway, I created a "sandbox" of five countries. And I can change who owns the CoTs and all. All countries get 12 merchants per year due to having monopolies in CoTs that no other nation has knowledge of. The CoTs do not disappear early in the game, and the whole world is colonized to allow the many CoTs. All empty or experimental CoTs are in Europe. I can change religion, capital location, stability, etc. and see if there is an effect. All I can then say is that the effect is as if trade efficiency was x% higher or lower. I thought about my sandbox quite a bit so that I could do all the testing with it. It is good to have my previous tests too. I want to see if the 3% per trade tech level matches up for example. Or the impacts may be different depending on relative strength of other nations for example.

Anyway, I have a very controlled but artificial testing environment. Luckily I can make some comparison to a real world situation.

[Miozozny]

I think therefore I am. That is all I know. But really how do you explain the staying power of pagans in distant pagan CoTs while other better traders are consistently being kicked out? If it is not religion then there is something else at work. From my examination of my Oman trials, I did not see staying power of Oman in Alexandria where Mameluks has the same Shiite religion. So I am baffled for now.

Of course in Alexandria most traders are from non-shiite countries. In Zacatacas for example most traders are pagan. Maybe that is more important than the religion of the owner?