The issue of obtaining benefits through bots and twinks
Write this article led me to a dispute on the Internet, as well as the long-observed senseless struggle with bots in games and additional accounts created to play for the second character, to give an advantage to your main character. I must say that this problem occurs only in low-intellectual games. Moreover, we can safely say where this problem is - this problem arises only because the developers did not want, could not make a balanced and intellectual game. Just think, what will the new account help you if you play chess? Well, tell me, chess :) But the principle can be extended to all other games. Then I was asked:
In the game “Cyber ​​Development” that I am developing, this is realized, it may not yet be final, but to a significant extent. And for those who are really interested in the answer - read on ...
The main thing - no bans on bots and additional accounts. It is necessary to do so that obtaining benefits through a series of tricks is not economically viable. Let's start with the trade, it will open the door for us ...
')
If you make the characters could not just pass on something to another, then everything will depend on the time spent on each character. And finally, if you play for two characters, you physically can't spend more time than you would spend on one. Therefore, when you play for two characters, you will create two nedopersonazhey, and not one cool.
Principle number 1. Not everyone trades! Getting into the game, the character can not trade immediately between cities. The character can only buy goods on the market in the city in which it is located. But first he has no money. To earn them, he must produce something and then sell it on the market of this city. Already here many tricks will fall off. If you start an additional account, you start to produce everything anew and without advantages. But here you have accumulated the cherished $ 100,000 for a couple of days and you were allowed to trade, and even more, you still have an additional account with a new character and you want to give him $ 100,000 to develop a twink so that he can help the main character.
Principle number 2. Natural market! The easiest way to realize a natural market is not to sell it :) Only the players themselves trade. But there’s a problem, they don’t have money initially, which means there’s no money supply in the virtual world at all. In some games, the player is given a certain starting amount, and more often they make an artificial market, when money is given by a certain government that prints money. If you give a starting amount, there is already a risk: the player starts 1000 accounts, and if there is an instant market in the game, they transfer them to the main account. But we have neutralized such a primitive scheme by principle No. 1. For a player to have 1000 accounts, oh well, even 5 - it takes a week to develop these 5 characters to the level so that they can trade. So, having no sources of money supply, we cannot create a natural market. We are introducing a government that prints money. But here you could grab my hand and say that this is an artificial market. Not really. It will become artificial when the money will only be issued, and the main thing is not proportionally issued. We will do so. The product is made in the game for a number of minutes. And if the government buys at a price that is proportional to the time spent in the game, the price will correspond to how much the player spent on playing time. Those. in our example from the first principle, the player will, without a difference, either spend the main character time to earn $ 100,000 for the second character, or spend the same time immediately as the second character. This is if buying a virtual government, but if you manage to make a deal among the players, it's even better - this is within the rules of the game.
Principle number 3. Time is transient! Of course, when an iteration in a game takes a large amount of time, say from an hour, then the player has time to build an additional character and control them simultaneously. But if your action takes a few minutes, or a maximum of 20 minutes. So how often will you manage to switch between accounts? How many can you start at the same time?
Principle number 4. Let's do extra. characters of the game element (solvable twinks) ! And we will not just solve it, but make advantages for them. For example, if from a new account a character gets into a random city, and in order for the main character to have the opportunity to trade with him, some of them need to accumulate horses and move into one city. We will give this opportunity immediately and make it an element of the game, like “find a friend / colleague, whom you yourself will manage”. Thus, a significant part of the "cunning" will be brought to light, but whoever is stubborn, and wants to get the benefits, will suffer great inconvenience.
We go further in debunking the possibility of gaining an advantage if we manage several characters simultaneously. But we summarize what has already been done. To honest players, we were given an advantage over "players who want the advantages obtained in dubious ways" (somewhere I saw a good term for this - a bag browser ). To become a Bagouser, you need to honestly spend time to get the opportunity to trade. Moreover, between the characters must be torn, because you need to track the status of each in a short time. By trading in the natural market from the "government" you get the amount proportional to the time spent. What opportunities still remain with the bagouser? Sometimes the game needs large sums of money or goods that the main character does not have, but others have in total. There is another option when something is needed at a particular moment as a "spoon for dinner", and then you need to quickly throw everything you have from different characters to one. Let's see how to neutralize this.
Principle number 5. The prices in the market are driving up! The instant market is only when trading in one city, and between cities it is necessary to make orders and the delivery of goods is worth the time and money. But it is not profitable for tvinkas in one city. Cities differ in what resources and how easily they can be mined. And if there is one, the other may not be. Therefore, for successful development you need to produce and trade in different cities. Now, if you trade between your characters, you can’t bet at a low price - the government or another player will buy for relatively little money. Then the bag-goer goes to the trick, the twink exposes some nonsense at the inflated price, and the main character buys it. Thereby the money is transferred from the main character twink. The second stage, the main character at an inflated price exposes the product that he wants to transfer and Twink buys. The main character returns his money, and the twin has the goods he needs. Note that in order to be able to charge up prices, the bauser had to save up the required amount of cash in an honest way. In this way you can merge all the valuable desired character. But alas, it will not be possible to do it instantly, because transportation between cities will cost you something (i.e., if you transfer to one place, you will receive less than you did in different places) and it will take time.
It would seem that here it is a bag-auxer which has merged everything acquired by one's own hard work, and he should get an advantage. But note, he merged only what he himself produced and earned during the game. His additional characters after plumming fell in the ranking, but the total changes did not happen. It remains to understand, and whether it will give him an advantage, the collection of goods and cash from one character.
Principle number 6. One of the goals of the game is to build sustainable production! We will make it easier, what we did, will declare the goal of the game to be the goal of the game and give advantages to those who do not hide their add. characters. Highlight a special role in the game - artisans. And we will make them on the principle of bot-twinks. We will give the player the opportunity to draw up a production schedule for them for the gaming week. They will automatically create products without participation of the player, and will be able to transfer the result for money from the main character at high prices. Someone will supply them with medicines, some with food. And the fact that he was doing the baguser, only will be an element of the game. Moreover, again, honest players will get an advantage.
Principle number 7. And if we want to tighten the screws? And you can cover up the shop and trade at high prices. How? How does the seller find a buyer? One option, the seller exposes what he wants to sell. The second option, the buyer makes an order for what he wants to buy. Suppose we have implemented only the second option. If the buyer places an order at a small price - he will never receive the goods, because he is first bought up by the government or the retail players. Then the buyer will place orders at an inflated price, such as the government will not buy. But in response to such an order, he will receive 10 proposals. If you make it so that he does not have the right to choose from whom exactly he will buy, then he will never intentionally transfer the goods.
Probably it is clear that if in the last section we neutralized bagouzers, with the ability to receive benefits by dubious methods, simply by fairly well implementing craft and trade, then there may still be opinions that if bots are played instead of the person, this can be an advantage. But in the article Games as a simulation of reality, I already wrote:
Let us describe an example of the kind when a task may not be solvable for AI methods, and therefore for bots.
All products and methods for their preparation can be described by an ontological graph. In some cases, it can be quite complicated, i.e. tightly bound. Consider one example of obtaining a "hearth" and using it a more advanced tool (pick, ax, saw).
The beginning of the graph consists of some items that do not require anything but the efforts of a mining agent. Such items are highlighted in green - stone and flint. But now logs do not require raw materials, but permission for logging. Such permits, as well as buildings (carpentry workshop, kitchen) require money. Permits and buildings are obtained / erected once and, unlike raw materials and tools, are not consumed. Money can be obtained by selling some product on the market. The more complex the product, the more expensive it is. But first, the agent is forced to trade in simple products, since does not have the ability to create more complex.
On the graph you can see the same subject, but with a different identifier. The fact is that there may be several ways to obtain a product. For example, “1. Logs "and" 2. Logs ”, in the first case, nothing is needed to get logs, but there are few logs, and in the second, an ax is needed to get logs faster. Passing along the graph in the reverse order, we will be able to outline the following action plan: Logs (2) - Ax - Center - Iron ore (6) - Firewood (109) - Stone ax - Stone pickaxe - Logs (1) - Flint - Stone (63) .
In general, there are different receiving options that may take a different amount of time. A person (natural intellect, EI) intuitively quickly understands that in order to receive goods more quickly, you first need to create a more advanced tool, and not receive goods in inefficient ways. But a person, more precisely, his intuition rather weakly helps, if you want to choose between two rather complex options for obtaining. For example, what will take less time to mine logs, first get an ax, or get a pickaxe to improve iron ore mining? But it already depends on the specific numbers - the time of receipt, the quantity of the product obtained in the chosen way, the prevalence of raw materials in the city, the efforts of the agent (the cost of health, energy, vigor, maintaining thermal comfort of the body). But playing a person over time abstracts from the exact calculation, and understands whether or not these or other parameters really seriously affect the result.
And what we have at the moment in relation to algorithms of artificial intelligence (AI)? In order to be able to use any of the AI ​​algorithms here, we will have to build a so-called. evaluation function ( synonyms : fitness function, utility function).
If the ontological graph would not have different ways of obtaining products, then it would be enough for us to use the so-called. direct and inverse logical inference (Chapter 7) . But alas, this is not enough. Between the different ways we need to choose. And here we have to form an evaluation function. I emphasize the evaluation function created by the developer of AI for each specific task, based on its formalization.
The question arises that if, in the case of a logical conclusion, we know how it is more convenient to represent knowledge, namely, everything can be reduced to an ontological graph, then in order for the fitness function to be built not by the developer, but by the AI ​​algorithm on the basis of ontology, we do not know which way is best. represent knowledge. It is possible for the formation of the evaluation function also makes sense to build a graph that will describe the specific numbers that affect the choice of the agent, and therefore the formation of the evaluation function.
So the creation of such a graph and its use for the formation of the evaluation function is the main problem for the current state of AI. And this is not about creating a private graph, but about a methodology for representing knowledge necessary for the formation of an evaluation function in the form of a graph (let's call it an evaluation graph ).
Finally, we will show how to try on natural and artificial intelligence. You just need to enter a certain index of success in the game. If this index reflects the amount of profit for the time spent by the player on the production of goods (maximizing profits while minimizing time), then winning the game with “brute force” according to the principle “you play more - you get more” will not work. And then programmers can engage in sports programming, competing, trying to make an intelligent bot, but players can simply show that natural intelligence is stronger, thus conducting a kind of Turing test.
In addition, this kind of success index can be useful if for some reason you do not want to tighten the screws in the fight against alternative accounts. According to the dynamics of this index, it is easy to understand who is sponsoring whom, perhaps this may occur because of not having the ability to trade, but then the legend is easy to introduce in the game that “a merchant loses a license when trading is not skillful”, but thus bagoiUrs will constantly lose licenses to trade . And note, we stop fighting at a low level by tracking the IP of the player, but just make the game intellectual and it automatically eliminates such problems.
How can one make a game in which two characters have no advantage over one?
In the game “Cyber ​​Development” that I am developing, this is realized, it may not yet be final, but to a significant extent. And for those who are really interested in the answer - read on ...
The main thing - no bans on bots and additional accounts. It is necessary to do so that obtaining benefits through a series of tricks is not economically viable. Let's start with the trade, it will open the door for us ...
')
The key is to trade
If you make the characters could not just pass on something to another, then everything will depend on the time spent on each character. And finally, if you play for two characters, you physically can't spend more time than you would spend on one. Therefore, when you play for two characters, you will create two nedopersonazhey, and not one cool.
Principle number 1. Not everyone trades! Getting into the game, the character can not trade immediately between cities. The character can only buy goods on the market in the city in which it is located. But first he has no money. To earn them, he must produce something and then sell it on the market of this city. Already here many tricks will fall off. If you start an additional account, you start to produce everything anew and without advantages. But here you have accumulated the cherished $ 100,000 for a couple of days and you were allowed to trade, and even more, you still have an additional account with a new character and you want to give him $ 100,000 to develop a twink so that he can help the main character.
Principle number 2. Natural market! The easiest way to realize a natural market is not to sell it :) Only the players themselves trade. But there’s a problem, they don’t have money initially, which means there’s no money supply in the virtual world at all. In some games, the player is given a certain starting amount, and more often they make an artificial market, when money is given by a certain government that prints money. If you give a starting amount, there is already a risk: the player starts 1000 accounts, and if there is an instant market in the game, they transfer them to the main account. But we have neutralized such a primitive scheme by principle No. 1. For a player to have 1000 accounts, oh well, even 5 - it takes a week to develop these 5 characters to the level so that they can trade. So, having no sources of money supply, we cannot create a natural market. We are introducing a government that prints money. But here you could grab my hand and say that this is an artificial market. Not really. It will become artificial when the money will only be issued, and the main thing is not proportionally issued. We will do so. The product is made in the game for a number of minutes. And if the government buys at a price that is proportional to the time spent in the game, the price will correspond to how much the player spent on playing time. Those. in our example from the first principle, the player will, without a difference, either spend the main character time to earn $ 100,000 for the second character, or spend the same time immediately as the second character. This is if buying a virtual government, but if you manage to make a deal among the players, it's even better - this is within the rules of the game.
Principle number 3. Time is transient! Of course, when an iteration in a game takes a large amount of time, say from an hour, then the player has time to build an additional character and control them simultaneously. But if your action takes a few minutes, or a maximum of 20 minutes. So how often will you manage to switch between accounts? How many can you start at the same time?
Principle number 4. Let's do extra. characters of the game element (solvable twinks) ! And we will not just solve it, but make advantages for them. For example, if from a new account a character gets into a random city, and in order for the main character to have the opportunity to trade with him, some of them need to accumulate horses and move into one city. We will give this opportunity immediately and make it an element of the game, like “find a friend / colleague, whom you yourself will manage”. Thus, a significant part of the "cunning" will be brought to light, but whoever is stubborn, and wants to get the benefits, will suffer great inconvenience.
We go further in debunking the possibility of gaining an advantage if we manage several characters simultaneously. But we summarize what has already been done. To honest players, we were given an advantage over "players who want the advantages obtained in dubious ways" (somewhere I saw a good term for this - a bag browser ). To become a Bagouser, you need to honestly spend time to get the opportunity to trade. Moreover, between the characters must be torn, because you need to track the status of each in a short time. By trading in the natural market from the "government" you get the amount proportional to the time spent. What opportunities still remain with the bagouser? Sometimes the game needs large sums of money or goods that the main character does not have, but others have in total. There is another option when something is needed at a particular moment as a "spoon for dinner", and then you need to quickly throw everything you have from different characters to one. Let's see how to neutralize this.
Principle number 5. The prices in the market are driving up! The instant market is only when trading in one city, and between cities it is necessary to make orders and the delivery of goods is worth the time and money. But it is not profitable for tvinkas in one city. Cities differ in what resources and how easily they can be mined. And if there is one, the other may not be. Therefore, for successful development you need to produce and trade in different cities. Now, if you trade between your characters, you can’t bet at a low price - the government or another player will buy for relatively little money. Then the bag-goer goes to the trick, the twink exposes some nonsense at the inflated price, and the main character buys it. Thereby the money is transferred from the main character twink. The second stage, the main character at an inflated price exposes the product that he wants to transfer and Twink buys. The main character returns his money, and the twin has the goods he needs. Note that in order to be able to charge up prices, the bauser had to save up the required amount of cash in an honest way. In this way you can merge all the valuable desired character. But alas, it will not be possible to do it instantly, because transportation between cities will cost you something (i.e., if you transfer to one place, you will receive less than you did in different places) and it will take time.
It would seem that here it is a bag-auxer which has merged everything acquired by one's own hard work, and he should get an advantage. But note, he merged only what he himself produced and earned during the game. His additional characters after plumming fell in the ranking, but the total changes did not happen. It remains to understand, and whether it will give him an advantage, the collection of goods and cash from one character.
Principle number 6. One of the goals of the game is to build sustainable production! We will make it easier, what we did, will declare the goal of the game to be the goal of the game and give advantages to those who do not hide their add. characters. Highlight a special role in the game - artisans. And we will make them on the principle of bot-twinks. We will give the player the opportunity to draw up a production schedule for them for the gaming week. They will automatically create products without participation of the player, and will be able to transfer the result for money from the main character at high prices. Someone will supply them with medicines, some with food. And the fact that he was doing the baguser, only will be an element of the game. Moreover, again, honest players will get an advantage.
Principle number 7. And if we want to tighten the screws? And you can cover up the shop and trade at high prices. How? How does the seller find a buyer? One option, the seller exposes what he wants to sell. The second option, the buyer makes an order for what he wants to buy. Suppose we have implemented only the second option. If the buyer places an order at a small price - he will never receive the goods, because he is first bought up by the government or the retail players. Then the buyer will place orders at an inflated price, such as the government will not buy. But in response to such an order, he will receive 10 proposals. If you make it so that he does not have the right to choose from whom exactly he will buy, then he will never intentionally transfer the goods.
The principle “you play more - you get more” must be replaced with the principle “you play better - you get more”
Probably it is clear that if in the last section we neutralized bagouzers, with the ability to receive benefits by dubious methods, simply by fairly well implementing craft and trade, then there may still be opinions that if bots are played instead of the person, this can be an advantage. But in the article Games as a simulation of reality, I already wrote:
There are tasks that are not possible to calculate mathematically, programmatically or in general somehow with the desired optimality. So if such a problem is incorporated into the game, then no plan for "pumping", any bot will not be able to play better than a person with his natural intelligence.
Let us describe an example of the kind when a task may not be solvable for AI methods, and therefore for bots.
Ontological graph
All products and methods for their preparation can be described by an ontological graph. In some cases, it can be quite complicated, i.e. tightly bound. Consider one example of obtaining a "hearth" and using it a more advanced tool (pick, ax, saw).
The beginning of the graph consists of some items that do not require anything but the efforts of a mining agent. Such items are highlighted in green - stone and flint. But now logs do not require raw materials, but permission for logging. Such permits, as well as buildings (carpentry workshop, kitchen) require money. Permits and buildings are obtained / erected once and, unlike raw materials and tools, are not consumed. Money can be obtained by selling some product on the market. The more complex the product, the more expensive it is. But first, the agent is forced to trade in simple products, since does not have the ability to create more complex.
On the graph you can see the same subject, but with a different identifier. The fact is that there may be several ways to obtain a product. For example, “1. Logs "and" 2. Logs ”, in the first case, nothing is needed to get logs, but there are few logs, and in the second, an ax is needed to get logs faster. Passing along the graph in the reverse order, we will be able to outline the following action plan: Logs (2) - Ax - Center - Iron ore (6) - Firewood (109) - Stone ax - Stone pickaxe - Logs (1) - Flint - Stone (63) .
Problematics
In general, there are different receiving options that may take a different amount of time. A person (natural intellect, EI) intuitively quickly understands that in order to receive goods more quickly, you first need to create a more advanced tool, and not receive goods in inefficient ways. But a person, more precisely, his intuition rather weakly helps, if you want to choose between two rather complex options for obtaining. For example, what will take less time to mine logs, first get an ax, or get a pickaxe to improve iron ore mining? But it already depends on the specific numbers - the time of receipt, the quantity of the product obtained in the chosen way, the prevalence of raw materials in the city, the efforts of the agent (the cost of health, energy, vigor, maintaining thermal comfort of the body). But playing a person over time abstracts from the exact calculation, and understands whether or not these or other parameters really seriously affect the result.
And what we have at the moment in relation to algorithms of artificial intelligence (AI)? In order to be able to use any of the AI ​​algorithms here, we will have to build a so-called. evaluation function ( synonyms : fitness function, utility function).
If the ontological graph would not have different ways of obtaining products, then it would be enough for us to use the so-called. direct and inverse logical inference (Chapter 7) . But alas, this is not enough. Between the different ways we need to choose. And here we have to form an evaluation function. I emphasize the evaluation function created by the developer of AI for each specific task, based on its formalization.
The question arises that if, in the case of a logical conclusion, we know how it is more convenient to represent knowledge, namely, everything can be reduced to an ontological graph, then in order for the fitness function to be built not by the developer, but by the AI ​​algorithm on the basis of ontology, we do not know which way is best. represent knowledge. It is possible for the formation of the evaluation function also makes sense to build a graph that will describe the specific numbers that affect the choice of the agent, and therefore the formation of the evaluation function.
So the creation of such a graph and its use for the formation of the evaluation function is the main problem for the current state of AI. And this is not about creating a private graph, but about a methodology for representing knowledge necessary for the formation of an evaluation function in the form of a graph (let's call it an evaluation graph ).
Success index
Finally, we will show how to try on natural and artificial intelligence. You just need to enter a certain index of success in the game. If this index reflects the amount of profit for the time spent by the player on the production of goods (maximizing profits while minimizing time), then winning the game with “brute force” according to the principle “you play more - you get more” will not work. And then programmers can engage in sports programming, competing, trying to make an intelligent bot, but players can simply show that natural intelligence is stronger, thus conducting a kind of Turing test.
In addition, this kind of success index can be useful if for some reason you do not want to tighten the screws in the fight against alternative accounts. According to the dynamics of this index, it is easy to understand who is sponsoring whom, perhaps this may occur because of not having the ability to trade, but then the legend is easy to introduce in the game that “a merchant loses a license when trading is not skillful”, but thus bagoiUrs will constantly lose licenses to trade . And note, we stop fighting at a low level by tracking the IP of the player, but just make the game intellectual and it automatically eliminates such problems.
Source: https://habr.com/ru/post/224833/
All Articles