Yea, this is really cool. Thank you to Maelstrom and Morda!!!

And yes, I AM KING OF BIZZARIA!!! HAHAHA I like Crateus too tho, very similar play-style, but a bit larger, so I might move in there also.

What are your net victories? It may make sense if you subtract your loses from your wins (total wins - total loses). Basically, it seems like a win is roughly +90 pts, depending on the score and RD if your opponent. A loss is roughly -90, again, depending on the score and RD of your opponent. 90 seems to be the default amount based on a score of 1500 and RD of 300 (starting stats). I haven't actually looked at te formulas, so if I'm incorrect, someone please jump in.

Here's an example, I've played 35 games on Biz, MacTob has played 49 games, but I have a much higher score than him because I probably have more net victories (wins - loses). Also, two of my wins were against MacTob, he won one against me (and is about to win a second one), and I have two wins against Tiger, needless to say, these wins bare a much greater weight than wins against people with lower scores.

In the end, I think these ratings are even less of an indivator of skill than the overall rankings because of the relatively few games played in each particular map. Just look at the RDs on even the most-played maps. Bottom line, if you think you should be at the top of a particular list, just challenge whoever's in the #1 spot, and beat the tar out of him. That'll prove you should be #1, and if you keep it up, you will be #1 very soon.

Similar play style! How do you get that? With so many allies and the countries so small and packed together, Crateus is a Brawl!

On the other hand, Bizarria has long borders with lots of trees, and very few countries, so its all about initial placement.

Actually, I agree that the two maps have a similar *feel*, though each for its own reason. On Bizarria, the borders are so long and the countries are so poor that a battle line can only work in two or three spots. Thus, it usually turns into tanks and mechs wandering around hunting each other, with the occasional LAV line thrown up to stall them. On Crateus, the countries are so small and the terrain is so open that guerillas have nowhere to hide, letting you simply conquer your way around a stalemate. Thus, it usually turns into tanks and mechs wandering around hunting each other, with the occasional LAV line thrown up to stall them.

So while they're very different on paper, Bizarria and Crateus both devolve into the same sort of tank/mech rugby scrum, with the difference being that Crateus has a MUCH higher kill count and a MUCH lower dependance on early positioning.


For the record, this is a very simplified version of what I think happens on these two maps...the gist of it, if you will.

My thoughts exactly Arty!

heh, I kept forgetting to post about this, but finally remembered.

Yes, I do agree with this, games on NP do take longer to complete than games on Bizzaria and are very time consuming to manage, so people who invest the extra time into those games should be awarded accordingly.

I agree with redfox. There is a sort of snobbery among some of the top players which says that the only real map is a large or medium map. It's a nonsense. Yes, the larger the map the greater emphasis on strategy and the less on tactics, and good for them if they enjoy that, but let's do away with this nonsense that somehow strategy is worth more.

What this boils down to in the end is they're worried that large maps take longer to finish. Well, that's the point of a statistical system like this (and its advantage over the MA system), in you don't need to play hundreds of games every day for your skill to be recognised.

Having an "overall" league, along with one for each planet and perhaps the ability to filter out players with an RD > 100, would make for an excellent rank IMO.

You really haven't left the MA score system mindset behind, have you? If you don't understand that LOTS of wins are not the be all and end all indication of skill, then, frankly, you're never going to understand what the glicko results are telling you.

oh yes...if i won 25 battles on that planet (perphabs losing 10) and Enforcer won 4 of course i suppose he is better than me...
very logical system system..next time Glickot will say you that some n00b who won one game over tiger and left the game forever is the best.

I just had an idea, why not list a player's MAN score and rank with the Glicko score, then we can view skill and experience together to get a better picture of a player's overall skills.

Think of it like this: say there are three players. Player A has had six matches and beaten Tiger six times. Player B has had twelve matches, beaten me nine times, and lost to me three times. Player C has had 22 matches, beaten the Easy AI 20 times, and lost to the Easy AI twice.

Obviously, Player A is the best. The Glicko system would agree, placing him at #1 among the three.

Although I'm no Tiger, I'm also not a pushover, so Player B would be the most experienced. The official ranking system would agree, placing him at #1 among the three.

However, Player C has the most wins. Under your logic, the player who went 22-2 against the Easy AI would be the "best".

Artanis what you rambling...i was not speaking about AI games or about games against easy opponent.i would not even spend time playing on NP against anybody with rank less than major.and had beaten TIger on NP too and others were also colonels and generals.
so you example about games against AI does not fits here.

We apparently are working on different definitions of "skill". By "skill", I mean how difficult somebody is to beat, which is what the Glicko system tries to count. As you see in my example, official rank does not necessarily mean skill (Player B).

This is not to say that opponent rank counts for nothing, of course. The higher the rank, the more likely the opponent is to have a high Glicko (for example, playing against a Marshal guarantees an opponent in the top 6% of the Glicko standings), but there are exceptions. Just look at the "Elite League" standings, where 1 in 4 Colonels has a losing record. By your logic, beating Colonel grtbrit (19 wins to 52 losses) would be more of an accomplishment than beating Zippo, who has played Tiger twice and never lost!

This is also not to say that number of wins count for nothing, either. Every time you win, your Glicko ranking goes up, no matter how bad your opponent is. You could beat a house plant on Bizarria, and it would still raise your Glicko! What the Glicko system does, however, is make a victory over the house plant, or me, or Pitor count for less than a victory over Tiger.

Just an apropos: The game "Laser Squad Nemesis", which has much of the same structure as MA (turn based, two player, play-by-email-ish) uses the ELO rankning system divided into three divisions (Novice, Veteran and Elite) and everybody is very pleased with it. ELO is almost the same as Glicko if I'm not mistaken.

And in case I blabbered too incomprehensibly, what I meant was: "Glicko = Good".

If anybody wants to get a feel of how difficult a houseplant is to beat, send me a challenge. It amounts to the same thing.

The other thing that needs to be taken into consideration is the RD (ratings deviation). The fewer the games played, the higher the RD is for everyone. High RDs represent uncertainty of that player's actual score. So, while Mrakobes has a score of 1825 on NP, his RD is by far the lowest on that map. Another example is with the overall Glicko ratings, Zippo is #1 with only 14 games because he beat Tiger twice, and is undefeated. However, his RD is double Tiger's, so there is much uncertainty whether he is truly skilled, or just lucky.

Glicko is based on ELO, and has been proven to work with time. Since we are still at the very early stages of this game, I imagine it will take some time for the truly skilled players to make their way to the top. I think eliminating extremely high RDs is a good idea. But again, the best way to prove you're the best on a particular map and increase your score would be to beat the people at the top of the list. If you really are the best, then you should be able to defeat the other top players consistently.

So challenge away, and let's see who comes out on top.

EDIT: Was trying to see if it would add my sig, think it will for my next post. In case you're wondering, my avatar is Talim from Soul Caliber 2, because she's my fav character and she kicks ass.

I've checked my list of battles. Both my games with Zippo was on Antarcticus (for both sides). I've sent new challenge and I hope we'll fight again

How the RD is supposed to work: Basically it is the deviation of what your actual score could possibly be. The Lower your RD, the more accurate the score is. So if you have a score of 1700 with an RD of 100, your actual score could fall anywhere between 1600 and 1800, and so with more battles you will find your score stabalizing at where your skill at the game would put you.

The RD is also used to increase the score won or lost in a match. If you play against someone that is more established with their RD, it affects your rating more. So if you beat Tiger, with his high score and low RD, you will obviously prove yourself to be a good player, and your score bonus will reflect that.

On the other hand, if you are beat by a weak player with a low RD, you will lose a lot of points. That player has proven that they aren't very good, and by losing to them, you show you aren't as good a player.

Losing to someone with a high RD does not hurt nearly as bad, as they have not established how good they are yet. They have had no chance to prove how good or bad they are.

In short, the best way to improve your score is to play and win against the people with the highest score and the lowest RD. But by the definition of how Glicko works, these people will be the hardest to beat.

the meaning of any ranking system is - the defining who is the best.
as i see the clicko ranking does not do this job - ranks there matter little because you have to consider all that RD maths.(sounds quite annoying to me).
We need the strict definition who is the first who is the second but instead we hear again "consider RD blah blah blah"
fact is - the clicko ranking places somebody who play just a few games and win over someone who play alot win alot but still lose sometimes
if you play alot you can not win all the time - you can be very good player but still there are some situations with deployment where you just can not win...even Tiger somehow faces the situations where he has no choice but surrender
in clicko penalties for losing are VERY high so high that they actually can pull down the best player who lost a few games by accident.
such system may be meaningful for a game where isnt ANY random but in MA the secret ally position matter alot
the current buildt-in MA ranking (as it is not very good) but however it inspires the players to play more.
this "glicko" system actually NOT inspires to play the game at all.

That's true, this game has a random element that Glicko was not designed for. I nearly got beaten by a rookie because he was PL and had a good ally setup. Imagine what that would do for his Glicko ranking.

I agree as well, there is definately a random element involved, and this has happened to me a lot as well. The only solution is to (following Tiger's example) re-challenge the person who beat you, I usually ask for best 2/3, and since I lost, I choose PL. I can usually win on the rematch as PL, and if you do, you've effectively cancelled out the loss because after their first win, their score will be higher, and their RD will be lower, so you should get at least part or all of your points back.

Then again, if you get unlucky a second (or third) time, then you're just screwed.