# How to implement a Counter-Strike rating system in a blockchain protocol, and why? - Kajetan Olas

- Channel: [ETH Warsaw](https://streameth.org/eth-warsaw)
- Date: 2025-10-07
- Duration: 18:36
- Watch: https://streameth.org/watch/yt-NO41j7qbwVM
- YouTube: https://www.youtube.com/watch?v=NO41j7qbwVM

## Description

What do Counter-Strike, chess, and DAO projects have in common? One key element is the need for reliable rating systems that can accurately assess user reputation and skill levels. Discover how methods long used in MMO games can enhance the blockchain economy of your project.


🧜🏻‍♀️ ETHWarsaw is a series of educational and entertaining events for an active community of blockchain builders, developers and enthusiasts with focus on Ethereum-related tech. Once a year, we organize a large conference and hackathon for the community in the center of the Polish capital with speakers from the best web3 projects and participants from all over the world.

Follow ETHWarsaw on social media for the latest updates!
X (Twitter): https://twitter.com/ETHWarsaw
LinkedIn: https://www.linkedin.com/company/ethwarsaw
Telegram chat: https://t.me/joinethwarsaw
See you all at our events in Warsaw 🙌🏻

## Transcript

how to implement a Counterstrike rating system in a blockchain protocol and why all right so thank you very much for coming uh and uh yeah that's what I'm going to talk about and this uh uh this I would say lecture is not going to be so much about the specifics of the country stri raing system but first we're going to talk about why do we even need sprting system in a blockchain protocol because uh the egalitarian approach is very popular there is a common common belief that it's good that everybody has the same vote but that's not the way democracies uh work in real life in real life uh people get Reed uh in relation to how much they contribute to society to their communities and uh that's what the rating systems are all about about assessing uh how much somebody contributes uh for to others and uh for example what will do have so uh first uh why do we display League of Legends because uh I think that's a good representation of uh what the rating system may represent uh they don't necessarily have to be very like complicated mathematical models u in the case of Leo flr or Counter Strike or any other MMO game um they are simply displayed as those divisions right silver diamond gold and so on so that everybody can so that everybody can U easily tell what skill does the other person have um so uh like I said uh the rating systems are all about assessing uh user skills uh and reputation uh reliably and uh it turns out that uh in the context of video games there were some pretty significant advancements uh in how those systems works just because the business needed it so in the case of like csgo Le Legends uh like uh other games made by Microsoft uh uh these systems are really really Advanced and we can actually apply them to blockchain and especially to uh to D so that uh people uh get rewarded for example for uh proposing meaningful projects uh for uh in the games of gamey uh for uh playing with others and uh let's look at an example uh why would we need even uh why would we even need a rating systems in governance uh well because currently most DS are purely based on how much token uh does somebody have right uh if you have uh 1,000 tokens then you have 1,000 votes um and it turns out to not be a perfect approach because uh that's a rather old report from chain analysis uh we can see uh they studied 10 major DS uh and it's not like they chose 10 that those are uh from what I understand uh the top 10 at the time so that's like uh uh so it's not like their job uh and uh in there you can see that uh 90% of all governance tokens of all votes were held by less than uh 1% of um stakeholders of of users so that's uh in the case of democracy or any other uh form of government in real life we could say that if 1% of people f 19% of all power then that's not a very good system and that's exactly the way things are in web free right now so let's look at the case of Unis swap um so that we can see why that's pretty bad in Practical example uh in 2023 uh probably some of you heard of it uh the unitab U submitted a proposal I mean one of the um one of the stakeholders it is uh and proposed that the Unis swap Launches on B&B blockchain using Wormhole Bridge uh and uh one of the major stakeholders the Andreas and horovitz Venture Capital funds um oppos it very much uh they actually have uh 4% of all tokens on you swap even though they said they had 1.5 of all tokens um and uh there was uh you know that's pretty funny tweet but um I think it's actually very accurate and uh later CZ from binance responded to it series analytical FS responded to it and um you know that's uh that's pretty much uh the nail and um in there you can see that uh um that uh the andresson harit for for that example controls 4% of all unisoft tokens that's a lot and uh they can single-handedly uh push most of the proposals that they want um so uh that's the case of unit which is the largest um Dow if I'm I'm not mistaken right now and uh the same applies to for example appcoin but even even more so because in the case of appcoin um the the Insiders control more than 50% of all all tokens so that's just decentralization is just a blood word and um there's are actually I think pretty smart ways to fix this so that people who are actual users of the project can have more more say and can decide uh what direction we want to go um so um that's another example but I think there's no need to go into it um how do we fix it so uh one way to fix it is to actually uh have a semi- reputation based system which is based on uh uh on the rating system so uh the way it would work is that for example in the case of gaml because I think that um the most uh that the best uh field where we can apply it uh we can uh have a multiplication factor which applies to tokens somebody has so for example if the Le League of Legends or Counter Strike at the token then uh they could have a d they could vote with the token but uh actually they could Implement also a making multiplication factor that for example if somebody is very good player uh like two standard deviations above the norm then we can uh for example multiply the number of their tokens by let's say two to the power of two right or pre standard deviation is sub the norm then two to the thir power right so um and obviously there would be a cup on how many tokens Can Be can be boosted this way so that uh funds don't have smar who who can um boost their votes so to speak um okay so uh now the Practical example of how we can uh how you can implement it or think about implementing it in your own D in your own project so this can be implemented in um any project that has uh that has some form of uh experience points that has some form of writing points uh and the way it would work is uh first you identify that area where players can compete with each other and it doesn't have to be like a very Pierce competition just has to uh be some kind of representation of how engaged somebody is and how much more engaged that person is than other community members because uh that's the uh I would say U objective way to tell if somebody should have more governance V votes than others um because that person is just more active in the community so if we are I don't know a developer who contributes who makes a lot contributions on p c then we can recognize it if that the of like the competition so then uh we can choose the rating system so here I very specific there are uh many rating systems those are just the three I would say most popular um and ELO is the most most basic one um so ELO is using chess in um yeah mostly in chess today but a lot of games even those Advanced ones use modifications of the ELO system so that uh uh so that you know some like correction factors are calculated using neural n network predictions so um for example uh yesterday I talked with a guy who makes a Counterstrike prediction systems uh like the tech cheating and uh he actually uses just the elos system but he added a lot of like factors that are calculated with AI to make up for example if there's a large chance that somebody's cheating then um that is included in it so below is the most cost isable one I would say um Lio is uh an improvement upon ELO in that sense that it include the time correction factor so it just means that if somebody doesn't contribute for a long time and then play then uh that person um has more rating variance so they can uh lose points more quickly if they play badly or they can uh gain them uh more fast because we are not so sure about their skill um and the true skill is one that's actually propretary system so um I'm not sure it's good idea to implement it because um it comes with having to obtain license from Microsoft if we want to use it legally and um it's used in uh games such as uh for example rocket League um yeah yeah I think it's rocket league in daa and uh the difference between truth skill and other system that truth skill has uh like Network Nal networks and deep learning and so on built by default so Microsoft develop these standards that they how developers should implement the system so that it's actually good for games like for example if we have a game with in which there are maps then we can say that if one of the players gets gets worse map or Worse side of the map than the other um then we can and uh for example uh make it uh have less effect on his rating because he has less chance to win so uh that's the second point in implementing this the rating system the second is uh that's uh that's the mo I would say the the one of the simpl step even though somewhat may seem somewhat mathematical on the presentation because it just means that when we have those players and they rent their ratings then we just uh say okay how many players are like uh what rating is required to be in the top 10% of players top 20% and so on uh mathemat mathematically we would say uh how would uh uh what rating does someone need to have would be two standard deviation sub amount standard deviation and so on um and uh that's required in other of this rating systems and then uh that's like a simplified example of uh you know what would we want to do if we want to make the most basic uh semi reputation B system so when we have uh those uh probability distributions and deviations worked out uh then we can say that uh some about this holding power is just tokens times the uh some multiplication factor for example like 2 to DX uh and uh where X is the number is the number of standard deviation of the norm uh in that player group in that community so that we can see how much somebody contributes and uh obviously this subject to modification a lot of modification if you want to defy but that's just an example so um if we have that player Bob who has a rating of 2,25 so uh if we go back that's 1.5 1.5 standard deviations above the norm so that player is pretty good for example in I don't know contributing on G or in u of game f the your game uh or in the games of like telegram clicking games in just clicking the the button right so he's better than most players um then we can uh make his uh vote more significant because he's engaging our platform we want him to uh have uh you know a lot of U governance U decision so um yeah and one thing to also mention here is that um we we want to have some kind of C like we don't want somebody with 1% of stake in all of our tokens to be able to multiply all those tokens just because uh then we would still have all those VC FS who just hire some students who plays uh the game for them or uh they just have an employee who who does this um but we want to for example C it to I don't know um tokens worth like $15,000 I think would be you know some um an amount that we would have more or less but it's not like we would make it 0.1% of all Supply but more like a couple couple thousand dollars and so um yeah that's all I'm uh happy to answer all your questions if you have some item of talking with you and uh feel free to uh join us uh on Twitter um we are uh I'm from PD company so we make uh we make games mini games that teach uh blockchain programming um so we actually use uh and develop some of these uh algorithm that I talk about uh we will uh in a few days publish a white paper which will describe this u in more detail so if you want to have some like um inspiration on how to do this in practice uh with you know more detail then you know feel free to um follow us thank you that's all does anybody have any questions at all sure like uh rating systems are in general kind of well understood so my question is different how do you measure uh like the impact how do you provide the input for the rating system in the case of like da or any kind of project like what a possible approaches here like I I can think of uh asking all of my employees like to rank each other pair of employes in the company right who is giving a bigger impact right like it's probably gameable right so what can be done here yeah sure so in the case of like uh un swap right because this obviously the best application the case of gaming I mean blockchain gaming game f and so on um I think that's one way to do this would be for example to uh when we have like uh for this Unis swap uh Dow block right we have the O proposals where people can vote for them with their tokens and uh if they reach a certain cap then they go through to the next B and so on so we could actually not uh you know uh just uh put the idea in the trash it doesn't reach at certain cup but actually reward is outter with uh some um reputation points and so um we can uh like make some you know anti U out to puts um anti scam uh protections so that if we see that some wets are connected and there are like pretty good open source software that can see if well are connected v v for something um then uh we can say that if somebody has any votes like that don't come from that that very person then that person should uh get some number of reputation points and those reputation points uh would be uh you know would pass with time uh people would lose them it's not like they would reset after each but um if we see that uh that person is less active in relation to everybody other then they lose them so uh that's how say just ranking liking posts with their tokens and reing everybody get some kind of Engagement uh well I have a question like uh when you say like rating this medivation all these things I was wondering if the rating Works similar like a trust algorith you know you what I mean like people tend to put impact between themselves in the D and yeah yeah so actually uh the way we we try to solve it is make the uh make the algorithm op open source so that there's no like Facebook that uh you know this Black Box nobody understands what is there and uh it's actually kind of destructive but you know we're uh I think it it can be done with the web foros so that all algorithms are open source especially is those simpar ones like ILO which actually there's research that shows that even ILO uh shows similar results to all those complicated more complicated algorithms like TR skill and gleo so it's not like it's worse than them it's just simpler uh so it can be very easily made open source and I think the the strength in implementing those rating systems is that uh you can actually you know spend the day and make some drafts of how we could work in um uh in many gamey projects like especially those on tgram with like da we said uh this would be a bit more complicated because their like simple L system will will not suffice I think but uh but yeah I think we can make it it's it can be made open source and thank you