Jump to content


Match Maker, Match Maker, Make Me A Match

MM Match Maker Devs Patent

  • Please log in to reply
136 replies to this topic

Perrigrino #101 Posted 04 January 2019 - 09:34 PM

    Senior Master Sergeant

  • Member
  • 6031 battles
  • 323
  • [OWSS] OWSS
  • Member since:
    10-02-2013

View PostblindfoId, on 04 January 2019 - 12:55 PM, said:

The ratio is desired and the MM aims to get to this ratio as close as possible (that how it works and is written in the article) but it may fail to do so. Please understand, I don't say that MM system is ideal. Not at all. It needs work and it is being looked into constantly. The article describes how it works at the current moment, and it is not perfect system but it is explained as plain and clear as possible and reflects the reality of its principles. 

 

I also do face some frustrating matches with seemingly unfair set ups but it doesn't mean there's some hidden mechanics or predetermined outcomes. Explanation was requested by our community and was given. I perfectly know that I cannot reassure or convince you or make you believe in the explanations but I'm still trying. It's wrong to say that we do not give an answer. The answer was given in the article and it is purely your choice either to believe in it or not.

 

For example, one of my last matches ended up like this:

Spoiler

 

Was I satisfied? Clearly not. But it happens to all of us :unsure: 

 

Again, the MM needs adjustments. Still the principles of its work was described in details and there's nothing over them - hidden or implied. 

 

Whole heartedly agree, and empathise that you are the messenger, tasked with a heavy load. 

 

I think the words "determined", "predetermined", are used synonymously and with "rigged", "fixed", and "cheat", and this needs to stop, as it only adds to the confusion. Understanding and trying to encourage change for the betterment of the game is the goal here.

 

  1. One of the main issues is about "HOW: match maker sorts players and puts them into match ups. And "HOW" this can be refined.
  2.  For "how" and "why" PRNG is "determined" see random.org- has more to do with the data set/distribution and, indirectly very little and likely next to nothing with battle outcomes being "predetermined", "rigged" "fixed" etc.,- if the model is imperfect, the flaws will appear. It has been admitted by the devs, that the model/algorithm/ (mm template ?) is not "ideal". 

 

3. "This strikes me as a system that’s only going to cause the bulk of players, who already seem to have a hard time grasping probability, to have even less realistic expectations the next time they encounter a game with true percentages.I mean, seriously, every luck-based game (Blood Bowl, XCOM etc) seems to have a seething forum full of people that think the AI must be cheating because 90% should mean a guaranteed hit. Speaking as an educator it’s legitimately horrifying.

When it comes to games is that players already know games already “cheat”. Games raise health and damage to create tougher enemies, give the AI side bonus or even infinite resources, read position and/or state data directly to determine AI reactions, etc… With that knowledge, it isn’t unbelievable to suspect that the game is also fixing random rolls. Funnily enough, games that actually rig their random rolls are most likely attempting to favor the player.

There also still seems to be a general distrust of pseudo-random number generation. People are more willing to trust physical dice rolls than they are a game’s RNG (and even then people will question whether the dice are fair.) Play a board game where one person makes every roll again and again against overwhelming odds, and people chalk it up to luck (particularly if they keep swapping dice around). Play a computer game where the same thing happens, and people believe the game’s RNG is busted or the game is cheating. (Differences are that you can’t actually see what the computer is doing the way that you see a physical die being rolled, and that you can’t “swap dice” if someone does question fairness.)

And, to be fair, the RNG in a game at its best isn’t going to cover the full range of possibilities. In reality, something that has a 90% chance of working might work 1000 times in a row, but it could potentially be impossible to have such a run. It doesn’t help that some games have worse designed RNGs than others, and some have particularly poorly implemented usage of an already iffy RNG design.

Two other big psychological issues are that success/failure is often binary and games end up in situations where the percentage chance doesn’t necessarily match what the player would believe to be realistic. These two can feed off of each other, such as when your trained soldier fails to even scratch the unarmored foe that he is standing right next to. In a real-time game, that kind of shot wouldn’t be guaranteed, because the target might through luck or skill dodge the shot. In a turn-based game, the player seems the foe just “standing” there are point-blank range, so you already have the player having a different perception of reality than what the game has. And then the player misses, and … nothing… The player gains nothing, because either he would hit or miss, with no middle ground. He won’t inconvenience the opponent (who realistically would have been spending time moving around while under attack), and he can’t scratch the opponent unless he hits (where a hit in turn may trigger a second check of the RNG for how much damage the hit delivers. So even a hit could itself turn into a scratch.)

And then you top all this off with a system that puts a lot of importance on a limited number of random checks, so things don’t get to even out over time. A single check breaking the wrong way, much less a full turn’s worth of “bad luck”, can wreck a scenario.- :https://forum.rockpa...ations/13608/55

 

Sry SV, but Specialization does not equate to "determination"- has to do with balance, which is definitely an issue. If we have superior vehicles (specialized) as part of the recipe, but the sorter (MM) does not account for those ATTRIBUTES, then that will lead to an imbalanced game- and the optics are, what they are. There are other issues of balance, like stacking each side with uneven vehicle types/classes. Then there is the issue of Tier distribution. This MM was designed by humans and therefore by it's very nature flawed. The MM model presently used, sorts for the things described in the Dev Blog. Can it be made better?- YES!

 

To do so we need to have a better understanding of just how it works, so that we can focus on ways in which to improve. I have endeavoured to gather information for these dual purposes.

 


Edited by Perrigrino, 04 January 2019 - 10:07 PM.


Perrigrino #102 Posted 04 January 2019 - 09:46 PM

    Senior Master Sergeant

  • Member
  • 6031 battles
  • 323
  • [OWSS] OWSS
  • Member since:
    10-02-2013

And, elsewhere, Blindfold has informed us: 

 

"Thank you for being involved and interested in World of Warplanes! I passed along the request for the article similar to the one which described the mechanics of Match Making System, as soon as team will come out with article the news will appear on our official portal, so stay tuned! 

 

If one takes the time, they will see that World of Tanks Wiki has an abundance of information regarding MM, Mechanics, RNG, etc., that provides detail and clarity. I have provided some snippets here (above) for those who care to look. One thing for sure, were getting somewhere.

 

And thank you Blindfold.



Perrigrino #103 Posted 04 January 2019 - 10:24 PM

    Senior Master Sergeant

  • Member
  • 6031 battles
  • 323
  • [OWSS] OWSS
  • Member since:
    10-02-2013

onward, look to the diagram on the far right. This is what I am driving at when referencing "determinism"- patterns/sequences are repeated. Full stop.:

 

Linear congruential generator

From Wikipedia, the free encyclopedia
 
 
Jump to navigation Ju
 

A linear congruential generator (LCG) is an algorithm that yields a sequence of pseudo–randomized numbers calculated with a discontinuous piecewise linear equation. The method represents one of the oldest and best–known pseudorandom number generator algorithms.[1] The theory behind them is relatively easy to understand, and they are easily implemented and fast, especially on computer hardware which can provide modulo arithmetic by storage–bit truncation.

The generator is defined by recurrence relation:

 

Period length

A benefit of LCGs is that with appropriate choice of parameters, the period is known and long. Although not the only criterion, too short a period is a fatal flaw in a pseudorandom number generator.

While LCGs are capable of producing pseudorandom numbers which can pass formal tests for randomness, this is extremely sensitive to the choice of the parameters m and a.[lease_clarify" style="background: none; color: rgb(6, 69, 173); text-decoration: none;; background: none; color: rgb(6, 69, 173); text-decoration: none;" title="Wikipedia:Please clarify">clarification needed] For example, a = 1 and c = 1 produces a simple modulo-m counter, which has a long period, but is obviously non–random.

Historically, poor choices for a have led to ineffective implementations of LCGs. A particularly illustrative example of this is RANDU, which was widely used in the early 1970s and led to many results which are currently being questioned because of the use of this poor LCG.


Edited by Perrigrino, 05 January 2019 - 10:17 AM.


Perrigrino #104 Posted 04 January 2019 - 10:29 PM

    Senior Master Sergeant

  • Member
  • 6031 battles
  • 323
  • [OWSS] OWSS
  • Member since:
    10-02-2013

had a devil of a time trying to get the diagram in the far right of the page to copy over. What it shows, the algorithm at work, repeating a sequence- creating a pattern given different variables..

 

"A benefit of LCGs is that with appropriate choice of parameters, the period is known and long. Although not the only criterion, too short a period is a fatal flaw in a pseudorandom number generator." -empahasis added.

 

https://en.wikipedia...ntial_generator


Edited by Perrigrino, 05 January 2019 - 10:10 AM.


Perrigrino #105 Posted 05 January 2019 - 10:20 AM

    Senior Master Sergeant

  • Member
  • 6031 battles
  • 323
  • [OWSS] OWSS
  • Member since:
    10-02-2013

View PostSenatorTH, on 05 January 2019 - 12:21 AM, said:

 

Then according to your thoughts if Joseph Goebbels would describe the principles of gas chambers work in details, leaving nothing to hide, all those millions of executed victims had to feel highly satisfied?

 

Please, your not helping. Blindfold is doing a very good job under the circumstances. And the association your trying to make is inappropriate.

Perrigrino #106 Posted 06 January 2019 - 02:46 PM

    Senior Master Sergeant

  • Member
  • 6031 battles
  • 323
  • [OWSS] OWSS
  • Member since:
    10-02-2013

interesting use of words- deterministic, random,

 

Pseudorandomness

 

A pseudorandom process is a process that appears to be random but is not. Pseudorandom sequences typically exhibit statistical randomness while being generated by an entirely deterministic causal process. Such a process is easier to produce than a genuinely random one, and has the benefit that it can be used again and again to produce exactly the same numbers, which is useful for testing and fixing software.

To generate truly random numbers would require precise, accurate, and repeatable system measurements of absolutely non-deterministic processes. Linux uses, for example, various system timings (like user keystrokes, I/O, or least-significant digit voltage measurements) to produce a pool of random numbers. It attempts to constantly replenish the pool, depending on the level of importance, and so will issue a random number. This system is an example, and similar to those of dedicated hardware random number generators. Even with all of this work, it is not truly random.

 



Perrigrino #107 Posted 06 January 2019 - 02:49 PM

    Senior Master Sergeant

  • Member
  • 6031 battles
  • 323
  • [OWSS] OWSS
  • Member since:
    10-02-2013

Measuring Skill Level and Optimizing Player-Matching Algorithms in Online Games

A major video game company wants to better match players in online game rounds. Dataset includes 1.5 million players of an online war game & individual stats from 150K play sessions.


The Wharton Customer Analytics Initiative is pleased to announce a truly unique data set from a major international video gaming company, which will allow unprecedented insight into the habits of online gamers. The data arises from play on a single multiplayer game and includes: game usage, game play, and historic behavioral data for 9.5 million users who played 882,000 unique game rounds. Complete longitudinal tracking is done for two cohorts of players, covering every round played by each cohort member, as well as all the players they ever played with, from product launch to March 2014.

The data sponsor believes that happy players will play more and buy more we want to help them develop new methods to test this hypothesis and develop operational improvements based on these findings. More specifically, the sponsor believes players are happiest when matched with others of similar skill level. Traditional skill-matching algorithms currently used in games do not take into account improvements in technology or detailed playing behavior that is available from today’s multiplayer games. The sponsor seeks next-generation skill assessment tools, as well as practical solutions to optimizing large-scale matching algorithms to reduce wait times for new games.

The project sponsor is open to other avenues of research, including but not limited to: social networking in online multiplayer games, online game behavior choices, or platform-based player segmentation.

Note: This Research Opportunity remains open for proposal submissions. Interested researchers can submit proposals online through the WCAI Archived Proposal Submission Portal. Researchers are encouraged to review proposal submission guidelines and FAQs before submitting their proposal. Additional questions can be directed to wcai-research@wharton.upenn.edu.


Edited by Perrigrino, 08 January 2019 - 07:33 AM.


comealong1 #108 Posted 06 January 2019 - 03:01 PM

    Command Chief Master Sergeant

  • Closed Beta Tester
  • 8351 battles
  • 704
  • [GW__S] GW__S
  • Member since:
    06-05-2013
Hmmm what a mouthful... A good read for sure.

Perrigrino #109 Posted 08 January 2019 - 10:24 PM

    Senior Master Sergeant

  • Member
  • 6031 battles
  • 323
  • [OWSS] OWSS
  • Member since:
    10-02-2013
 

 

forward,

 

Computational methods[edit]

Most computer generated random numbers use pseudorandom number generators (PRNGs) which are algorithms that can automatically create long runs of numbers with good random properties but eventually the sequence repeats (or the memory usage grows without bound). These random numbers are fine in many situations but are not as random as numbers generated from electromagnetic atmospheric noise used as a source of entropy.[9] The series of values generated by such algorithms is generally determined by a fixed number called a seed. One of the most common PRNG is the linear congruential generator, which uses the recurrence

 


 
 

Perrigrino #110 Posted 08 January 2019 - 10:59 PM

    Senior Master Sergeant

  • Member
  • 6031 battles
  • 323
  • [OWSS] OWSS
  • Member since:
    10-02-2013

Algorithm in a nutshell:

 

suffice to say, the MM algorithm is nothing but a fancy SORTING tool, used to sort players into matches. It is much more quicker and efficient and is completely dependent on what criteria is used- which is dependent on what the human(s) use to put into the construction of the Sorting Tool. We do not know how many MM are used, what kind of RNG is used by WoWP devs, or what Criteria is used, beyond Class and Tiers, but we are given information about the Template:

 

The algorithm is as follows:

Step 1. Flights are added from the queue, maximum one per team. At this stage, it only considers aircraft tier in these flights and ignores classes completely. If there is a single flight ...

Step 2. The matchmaker starts pulling single players from the queue. Firstly, it balances the class group ratios that appeared after adding flights. For example, if one flight included an attack aircraft and a fighter, while the other two players are piloting two “heavies”, the matchmaker will add a PvP aircraft to team A and a PvE aircraft to team B. When doing this, it will additionally consider the totals of aircraft tiers in these flights and try to equalize them with single players.

Once the initial balance is achieved, the matchmaker continues to add pairs of single players from the queue (one per team) with same-tier aircraft when possible. When doing this, it considers desirable ratios for PvP and PvE classes, which depend on the game modes. For example, in the Conquest game mode, a team may have up to 7 PvP-oriented warplanes and up to 6 PvE machines (not simultaneously). In other game modes, these limits are different because the ways to gain victory in battle are significantly different there. For instance, in the Escort game mode, one of the teams cannot have any PvE aircraft at all.

The current maximum number of players per team is 9.

When forming teams, the matchmaker is constantly adding up their aircraft tiers and comparing them, aiming to equalize the result. This is done to reduce the waiting time and make the teams’ chances of victory as equal as possible. When there is only 1 player per team in a battle, there can be no difference in their aircraft tiers. When there are 2 player per team, the matchmaker may allow a difference of 1 tier for them. In 3vs3 and more, a 2-tier difference is allowed. This is subsequently negated by adding bots.

Step 3. The matchmaker balances the teams using bots based on aircraft tiers and additionally makes their class composition as close as possible. It aims to make team aircraft tier totals equal and make the number of aircraft of each class (not class group anymore!) as close as possible to a desired ratio ...

that's it in a nutshell. There's a little more that provides some detail to the roster's ratio and some of the rationale. Ultimatley, it is the rationale that, in my OPINION is flawed. I repeat, this is my opinion, based on what I have read and my understanding of simple mechanics and human imperfection. 

 

Step 1: only Tier

Step 2: Class Group Ratio

Step 3: Bots added based on Tier/Class- to a desired ratio

 

Does anyone else see a problem with the rationale which went into the construction of the MM?



Perrigrino #111 Posted 09 January 2019 - 12:18 AM

    Senior Master Sergeant

  • Member
  • 6031 battles
  • 323
  • [OWSS] OWSS
  • Member since:
    10-02-2013

Blindfold, in all honesty you responded:

 

"The ratio is desired and the MM aims to get to this ratio as close as possible (that how it works and is written in the article) but it may fail to do so. Please understand, I don't say that MM system is ideal. Not at all. It needs work and it is being looked into constantly....

 

I also do face some frustrating matches with seemingly unfair set ups but it doesn't mean there's some hidden mechanics or predetermined outcomes. Explanation was requested by our community and was given. I perfectly know that I cannot reassure or convince you or make you believe in the explanations but I'm still trying. It's wrong to say that we do not give an answer. The answer was given in the article and it is purely your choice either to believe in it or not....

 

Again, the MM needs adjustments. Still the principles of its work was described in details and there's nothing over them - hidden or implied."

 

Thank you for your candid remarks. Ideas regarding the "things" that need adjusting, the direction in which the adjustments should go needs some input from the Players. Left up to "one" (more than 1?) dev (s?) and signed off by the director, does not BALANCE make. 

 

Suggestions to improve (adjust) balance have been provided (and please do not say - "impossible" - some of these presently utilized in the other WG Titles):

 

  1.  No Type differentiation- Same Type. Identical.
  2.   No Tier differentiation- Same Tier battles.-No More Tears.
  3.  Flight - matched with Humans/Bots as close as possible- configure AI to target Flight, knowing this players can act accordingly.
  4.  Class Group- need to match Type to Type. No more GAA in Lieu of Bombers and vice versa. See #2.
  5.  Dedicated regular quarterly review for quality of balance- not every 2 years.

Beyond MM:

  1.  Specialization- available to all, no need to grind to unlock. Available at stock configuration and continuing to Elite status- Specialization runs parallel while working toward Elite.
  2.  Gold Ammo removed. Revenue generation can be effected elsewhere- consumables/equipment parts that could require replacing. 
  3.  Maps- some asymmetrical maps may need further review based on above.

 

Some ideas, thoughts to help us in the direction toward a more balanced Match Maker, since it is being looked into regularly (?). (please adjust list as needed)

 

Let's make WOWP Great Again!

 

 

 

 

 


Edited by Perrigrino, 09 January 2019 - 12:27 AM.


trikke #112 Posted 09 January 2019 - 02:58 AM

    Captain

  • Member
  • 3626 battles
  • 3,484
  • [R-A-W] R-A-W
  • Member since:
    01-26-2012

i'm starting to think that it's just not possible to get a perfect MM...   

 

most seem to get fairly close, to me, anyway 


Spittoon says #smarterpilotswinmore

Perrigrino #113 Posted 09 January 2019 - 10:47 PM

    Senior Master Sergeant

  • Member
  • 6031 battles
  • 323
  • [OWSS] OWSS
  • Member since:
    10-02-2013

View Posttrikke, on 09 January 2019 - 02:58 AM, said:

i'm starting to think that it's just not possible to get a perfect MM...   

 

most seem to get fairly close, to me, anyway 

 

This is not about the pursuit of perfection. Please stop trolling.

Perrigrino #114 Posted 09 January 2019 - 11:27 PM

    Senior Master Sergeant

  • Member
  • 6031 battles
  • 323
  • [OWSS] OWSS
  • Member since:
    10-02-2013

 

 

from,  https://www.c-progra...definition.html

 

Algorithm definition

    Algorithm definition: Algorithm is a finite sequence of steps leading to the solution of a given task. This could be break down to the following rules:

  1. The sequence must have a starting point. This is the first step that will be done.
  2. Each step(action) must lead to another step or complete the algorithm. This means you can't have 'hanging' blocks that don't go anywhere.
  3. The task must be completed in a finite number of actions.

Good practices

    We have just touched some of them but, let's reveal them better.

  1. Define the task as clear as possible. Good task definition is half of the solution.
  2. Solve a single task at a time. Don't make your life harder by solving several problems at once. If you need to solve another problem, create a separate routine.
  3. The routine should do its job well. The result should be correct and complete.
  4. Handle all possible cases. This is especially valid for more branched solutions.
  5. If the input data is incorrect you should report the error and return no result. You should not return incorrect result.

    Now we know the algorithm definition, we know the best practices, it is time to look at one example:

    Choose how to go to work:

  1. Do you live in walking distance from work?
    1.1. Yes – walk to work.
    1.2. No – go to step No. 2
  2. Do you have a car?
    2.1. Yes – drive your car to work.
    2.2. No – use public transportation.

    You can easily see that the starting point is 1. You can also see that the result is different, depending on the answers. A very simple routine/decision like that is easy to follow, when described with text. But in reality algorithms are more complex. That's why flow charts are created. They are the universal way to describe visually any sequence of steps. If this is new stuff for you, I suggest you take a look at the flow chart and flow chart symbols lessons from the beginners programming tutorial.

 

Of course, this is just a very simple example. It could not be true for everybody – some people could ride their bike to work or even work from home. The example is just to show you, that various actions and decisions could be broke down to simple steps. Programmers do this every time when they create a new software. 



Perrigrino #115 Posted 10 January 2019 - 09:37 PM

    Senior Master Sergeant

  • Member
  • 6031 battles
  • 323
  • [OWSS] OWSS
  • Member since:
    10-02-2013

Found this in refence to a match making system that uses player profiles. What is interesting is the problems faced by utilizing such a system:

 

https://dev.dota2.co...ad.php?t=129665

 

Matchmaking Explained

Surprisingly not a suggestion or complaint thread. This is a short playdota article I wrote to explain matchmaking systems.


I won't pretend to be an expert, but I have an understanding of matchmaking algorithms. If anyone would like to correct any of my points, feel free to do so.





Your Matchmaking Score
-----------------------------



Actual matchmaking has nothing to do with "what kills you have" as that is a factor that isn't entirely affected by skill, and has negative results on players who aren't selfish. Which is why such an algorithm would actually be flawed.

Matchmaking algorithms are based on probabilities. If you are playing against players of an equal matchmaking score, you should have a probability of winning around 50%. The ideal of all matchmaking algorithms is to have you around a 50% win rate, because you should be placed against equally skilled players all the time.

On average, you out perform players below your current skill level, and perform worse than those above. Though you often will win against greater players and lose against worse players, as this is only an estimate of what "should" happen.

If you win against someone you should win against, your rating stays around the same, but if you beat someone above you, your rating will increase. The opposite is true if you lose.

The difference the loss or win makes is actually calculated by the difference between your skill levels. Beating someone way over your skill level changes your ranking much more, than the very tiny boost you'll get for beating someone only slightly over your skill level.

 

 

 

 



Perrigrino #116 Posted 13 January 2019 - 09:39 AM

    Senior Master Sergeant

  • Member
  • 6031 battles
  • 323
  • [OWSS] OWSS
  • Member since:
    10-02-2013

and for those that don't have the energy to read lengthy passages, or follow hyper-links, the following is provided: AGain, it relates to a Match Maker that utilizes Player Profiles, and the site discusses problems that these programers/developers face-

 

"Their is an additional variable that increases the magnitude of how much you lose or gain each game, which is used to make the first several games for new players worth much more toward their ranking than any other game, because it needs to get them somewhere in the system (though it's not an accurate ranking quite yet, this speeds up the process to approach it). This variable tends to ease out and eventually stop increasing the magnitude of each change when you have lots of matches played.

These matchmaking algorithms also tend to take how often you swerve away from your average probability of winning as a factor of how much your matchmaking score will change, to avoid huge jumps for players who have lots of wins against better players and lots of losses against players who are worse.

If we are talking about what determines your performance, it generally will be just how much you have won vs lost, and in a team game it will likely sum up the two teams player scores to determines the probability of who should win, and what impact that will have.

 


Many systems take into account how close each game was, though we do not know if this is a factor used in Dota 2's matchmaking system.

Some players like to party with friends of much lower level than they are. This is why you sometimes get players who are a lot lower level than you are in the game, and though the system won't jump this lower level player for high level wins if it already understands that players matchmaking rank, it could mess up if they are still in their first few games.

Smurf detection is it's own complicated issue, but we know for sure that the system checks if you completely pub stomped almost every one of your first several games, and if you did, it will skyrocket your matchmaking ranking. So high level players with new accounts will quickly leave the low level matchmaking pool. This is the only scenario where kills and deaths have any effect on your matchmaking rank.

The system can't grantee the outcome of a game even with what is supposed to be a very balanced match. Also, we have to consider that dota 2 games are also affected by picks, communication, and who is playing what hero, which is even MORE of a nightmare to judge, in comparison to determining the skill difference when both players are playing the same heroes/soldiers/have the same units.

 



Perrigrino #117 Posted 13 January 2019 - 09:41 AM

    Senior Master Sergeant

  • Member
  • 6031 battles
  • 323
  • [OWSS] OWSS
  • Member since:
    10-02-2013

and, continuing from that discussion:

 

Finding a Match
-------------------



When you enter matchmaking, you enter a localized pool of eligible players to join a game.

What pool you are in is determined by the region you selected (for example, US East, Europe West, etc...). If you selected multiple regions, you are in every pool you selected.

This pool lists you by your matchmaking ranking, and during the time you wait, it's going through players within a certain range of your skill level and placing them in a match lobby with you.





If it doesn't find enough players within that range the first time, then it will repeat the loop, except with a less restrictive range, meaning their is a larger skill differences between you and these players, than the difference between you and the players it has already found.







This loop repeats until it finds ten suitable players, and has balanced the teams between those ten suitable players found. Once you are in a match, you leave your pool(s) and connect to the game server.

So the skill difference between you and the players in your game is determined by what players are available, and even if their is a large skill difference, it's still better than you waiting in the matchmaking pool for hours, waiting for that perfect match.

Specific to Dota 2, the larger your party size going into a match, the lower the range of skill between you and your opponents it will tolerate. So you will wait a little longer for what will be likely be a better quality match.

So you are more likely to find a good match if you have a ranking with a lot of players, which is around 3500, at the time when the most players are on, and in the highest traffic region.

So unfortunately, this means that in the reverse scenario, if you are near the highest or lowest ranking possible, where there are fewer players, you play around 3-5am, and if you live in a region with the least players, you will have matches of lower quality, and longer wait times.

This is why some players decide to play in other regions than their own, because they want to avoid the tremendous wait times they get in their region. Some regions also have terrible internet, so they will get the same ping nearby as they would in another region anyway.

 

 

https://dev.dota2.co...ad.php?t=129665






 

 

 


Edited by Perrigrino, 13 January 2019 - 09:43 AM.


Perrigrino #118 Posted 14 January 2019 - 04:22 AM

    Senior Master Sergeant

  • Member
  • 6031 battles
  • 323
  • [OWSS] OWSS
  • Member since:
    10-02-2013

poking around afore-mentioned forum (see above) came across this, which gave me a whole new insight into BOTs and what is "scripted" for these creatures. Got me to thinking, maybe should keep in mind which BOTs do well, at which Tier..., and also, wouldn't it be interesting to have a Co-op battle mode with Super BOTs?- just an idea....

 

"

 
Chris C
Valve Developer
Join Date
Sep 2011
Posts
1,704

Upcoming changes to bot scripting

We've got a couple changes coming up that I wanted to make you guys aware of:


Dedicated Server support for bot scripts
We're starting to roll out support for running bot scripts on our dedicated servers. Much like custom games, we'll probably be selecting a few scripts at first to test the system, and then well likely open it up to the most popular bot scripts. This enables us to do a few things:
  • Coop bot matches vs scripted bots. We're not exactly sure what this is going to look like, but we'd like to do a better job of exposing scripted bots to the segment of players that enjoy playing against (or with) bots. Some sort of integration into the play menu is likely.
  • Bot-vs-bot matches. We'd like to publicize (and make spectatable!) bot scripts playing against each other, possibly by having some sort of presence on the Watch tab. Lots of details to still be worked out, but we think it would be entertaining for customers to get a preview of the top bots in action. Possibility of something like bot script MMR?
  • Bots that cater to more specific audiences. Having the highest-skill bots is a fun challenge, but we'd like to showcase bots that do other things -- specific challenges (dealing with the super-push bots is a good example), or bots that are more tailored to teaching new players the game, rather than mercilessly defeating them.


Expanding Workshop support
In the coming weeks we're going to move bot scripts to be more normal elements of the Dota Workshop. That means that scripts will have their own workshop page with images, videos, longer descriptions and update notes, comment sections, etc. You'll also be able to manage subscriptions from outside the game.



The goal with these changes are to better showcase your work, make sure players who are interested playing your bots can easily do so, and foster a development community of bot authors and their fans. "
 
https://dev.dota2.co...ad.php?t=277887

Edited by Perrigrino, 14 January 2019 - 04:28 AM.


Perrigrino #119 Posted 14 January 2019 - 04:55 AM

    Senior Master Sergeant

  • Member
  • 6031 battles
  • 323
  • [OWSS] OWSS
  • Member since:
    10-02-2013

And here's another idea for balancing Matches:

 

Re-divide the Tiers that MM considers for Match Ups. Consideration could be given to something like this:

 

Levels

  1. Tier 1-3
  2. Tier 4-6
  3. Tier 7-8
  4. Tier 9-10

and keep them separate. So no T3 vs T4, T6 vs T7, T8 vs T9, or

 

​Level

  1. Tier 1-3
  2. Tier 4-8
  3. Tier 9-10

 

with no Match up greater than +/- 1 difference in Tier. At the moment, T3 players battle in T4 and do not enjoy all the rewards associated with T4. The Daily Missions are divided into levels that also, somewhat, ignore the realities of the game (vehicle attributes for example: T3 vs T4 or T8 vs T9) and create match-ups that are "challenging", frustrating, and at times lopsided. Not saying these Matches are impossible, but given the amount of frustration felt, the number of times players "opt-out" of battle, need to look at all options when considering BALANCE going into the model of MM .

 

or, something else? Perhaps, no Tier differentiation? Matches based on TYpe and Class? So what one sees in the roster is a mirror image. Identical planes on either side. To shorten wait times, of course, fill with BOTs, as per usual.

 

And, and... while I'm on a roll, How about having a MAP with just TWO Sectors, no ADA- MORE LIKE CAPTURE THE FLAG. 

 

SAme vehicles, two bases, no ADA, conquest mode, spawning limited to SQuall Line.

 

Just throwing it out there.


Edited by Perrigrino, 14 January 2019 - 05:02 AM.


Captain_Underpants53 #120 Posted 14 January 2019 - 05:05 AM

    Captain

  • Member
  • 24068 battles
  • 3,482
  • [A-S-S] A-S-S
  • Member since:
    04-17-2017

Hail the bot overlords!

 

:playing:


MSgt, USAF, (ret)





Also tagged with MM, Match Maker, Devs, Patent

1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users