coan.net: Some knight fight tournys are set without a random start positon like the last team tournament and I see a advantage there if you study the game well. You can also end the game in several moves and if you play correctly and have a "good"set up with high numbers in your home side corners. I don't see a advantage with the colors at knight fight as long as you can see far into the game, and must always follow your opponents line of play. Now we talk about the advantages of the start colours but I think spider games and line games have the greatest advantage and white to some players like tenuki and masterx never loose white in spiders...or I think once in many years. I don't understand how starting on one particular colour can change things in knight fight as you can move to a different colour square on the next move, but that may be your secreat weapon, lol

Roberto Silva: In Knight Fight, I agree with lukulus - I would much rather be black.

With black, you can choose to start on the same color as White, which you can then limit the moves of white and "protect" spaces of high value and if you plan well, can then send them to a very low number and mess up their plan completely.

Since most players don't know of this "advantage" of being black & starting on the same color space, if I'm white - I will look on the opponents side of the board - find the highest value space, see what color it starts on - and then myself start on the opposet color, even if it's of lower value

Of course if you get 2 players who do little planning and just jump around until the end, then White might have the advantage of an extra move at the end - but I would still prefer to play as black.

pedestrian: 10% is still a relatively high probability. It means that one out of every 10 Brainking users will get such an unequal distribution or worse. You were just unlucky. As for the other question, yes, you can delete any game (non-tournament) within the first 2 moves, and it won't count for rating or statistics.

lukulus: Actually, white does have an advantage in Knight Fight due to starting the game. If black ends up unable to move, white will benefit from an extra move, which can mean up to an extra 99 points. Also because of starting, white always has first choice over which squares to delete. After several games that starts biasing the results a little in white's favor.

if you have someone on block can they still add you to there friends list?

if so shouldnt that option be removed

lukulus:

Yes, I suppose these percentages only give a very shallow picture of the possible color-biases. That was why I included several games, but maybe Knight Fight should have been left out.

Come to think of it, I've had a few games where people would take the challenge, then cancel the game immediately after. Is it possible (maybe only for paying members) to cancel a game when you got the "wrong" color, and avoid having the game rated?

pedestrian:Just a notice. Knight fight is a bit special. I would say black is slight better. 60% wins for white is just because lots of people dont know how to play properly.

I dont think there is some way how to cheat.

When I create new games in the "waiting games" section, I always choose random colors. When I accept waiting games from the list, I almost always go with those that have colors set to "random". So, in the long run you would expect the randomness to even out and I would have an equal amount of games with each color.

Now, as you probably know, some games here are very color-biased. As I browse through the games statistics, I find that these games give one side an overall score of approximately 60% or more:
Horde Chess (white 74%)
Cheversi (black 68%)
Maharajah Chess (b 64)
Mancala (w 64)
Pahtum (w 62)
Hasami Shogi (w 62)
Swap Five in Line (crosses 61%)
Battleboats Plus (w 60)
Knight Fight (w 60)
Tablut (w 60)
Swap Five in Line is a little complicated because you have the option to change sides after the first move, so I'll leave it out here.

That leaves nine game types that are very color-biased. When I look at the games I have played of these game types (always choosing "random colors"), they add up to 107 games combined. Of these I've had the "good color" 46 times and the "bad color" 61 times.

My grasp of binomial distribution is a little rusty, but if I get it right, the probability of such an unequal distribution should be less than 10%. So my question is: Is there some kind of trick than I'm missing (and that some of my opponents are not)? Is there some crooked way of side-stepping the randomness?

pauloaguia, Constellation36: Ah, i never understood what she wanted say with glicko and i didn't know that system. Well, now i saw my mistake and i must offer excuses to Abigaill.

Undertaker.: AbigailII just pointed out some flaws in the system you were proposing (which, by the way, I also think was preetty weak at some points)
Then took the opportunity to present once more the case for Glicko as a system that already covered some of thos flaws and even addressed the original this last problem you brought because in Glicko, in order to keep a good ranking, you must keep playing or it will decrease because you don't have many recent games...

I'd also like to see other rating systems in place at BK, but seems to me that this discussion is being missdirected because of some missunderstandings about some of the previous messages...

Undertaker.: Now, i see you don't defend it. So, do you defend anything or only know speak ill of all ideas (and create tournaments and more tournaments and bad tournaments lol)?

He does defend the Glicko system if you are not ill and can read. :-)

AbigailII: I give a suggestion and defended it and you spoke ill of my ideas, so i thought you defended the present system. Now, i see you don't defend it. So, do you defend anything or only know speak ill of all ideas (and create tournaments and more tournaments and bad tournaments lol)?

Undertaker.: You're free to have your opinion like me too. You defend bkr rating calculation and i defend other system...

Where the hell do you get that impression from? If you think the system I'm "defending" looks like BKR, then I don't understand what your problem with BKR is, because obviously in that case, you don't know what BKR is.

BKR isn't Glicko. Nor Glicko-2. BKR is what we have, and of which Fencer has stated repeatedly in the past is what we will have now and in the future. But don't get the impression that just because I see drawbacks in your suggestion that I defend BKR.

Undertaker.: To get to the top of any Pro field of any kind.You have to work very hard to do so.It is not done in winning 4 5 6 7 times,and to stay there you have to keep going until you retire not just sit there:))

"GERRY": Yes, unfortunately. I know some cases, but i think rating calculation is indifferent about that question...

Undertaker.: There is very good player's on this site that you can beat in a game & there is player's on this site that you can't beat no matter what you do.It's been like that since i joined in 2004 & i do not think it will ever change.

AbigailII: You're free to have your opinion like me too. You defend bkr rating calculation and i defend other system...

So, as you can see, there's a little problem here, between me and carl, because he wants to fight for the first place and i'm not interest in play more plakoto games, because i don't like of the game and it's not my fault if i'm in first place casually.

If you defend this system, so maybe you must get a solution for carl problem. Thanks and bye!

Carl: You're searching by the wrong fighter.
As i said before, i played 4 plakoto games, 3 of them were by tournaments with prize and one was to understand better the rules of game and this way prepare me for that tournament. My opponents had and have weak bkr's and i don't know how, but i have 2457 points. I don't like very much of plakoto game and i must only play more plakoto games by tournaments with prize or by team tournaments.

So, i'm not going play more plakoto games and accept invitations only because i'm in first place. There're others games as PahTum or Spider Linetris where i only played few games and i have bad ratings and i don't play them too, because i don't like very much of them. The reason is the same, and if you want to be in the top, you must fight to change this wrong rating calculation (in my opinion, of course). With my idea about rating calculation, i would only have few points, and would be more fair, as i tried to explain before.

To finish, your post, trying to take advantage of this discussion for i accept your invitation, was worthy of a failled oportunist. I'm disappoint with you.

Undertaker.: Did you already see that rating calculation in some game site for say if is good or bad? I don't think so.

Glicko you mean? It's used by FICS (free internet chess server), schemingmind.com and by the Australian Chess Federation, among others. I've played at both FICS and schemingmind. And I think it's also used by the ICS (internet chess server), but I haven't played there.

Undertaker.: You can make your plakoto rating more realistic by simply accepting one of the invitations i send you and actually playing the game! (Like in everything in life, to be in the top is need "fight").

AbigailII: Did you already see that rating calculation in some game site for say if is good or bad? I don't think so.

If i talked about that is because i know a game site where that system exists and it's very good. Besides, there's a similar system in Portuguese Checker Federation and in many others Checker Federations, so you say nonsenses.

With the knowledge i have, in this moment, about brainking and about many games that there're here, i can create a new account and i would be in first place in many ratings, playing 4 games with good players and winning, and then i would only play with weak opponents and would make 25 games, maintaining my first place. I'm not obligated to play against second person in rating, so it's easy to be in the top and maitain me there.

With the other rating calculation that i talked before, you will have many difficulties to do it.

Sorry, but you see BKR system in some sports? It would be funny see in Tennis, a new player make 4 games and with lucky (or not) win those 4 games and become the number one of the world ahahah...or see a new country like Timor make 4 games in football and be in first place in Fifa rating.

Like in everything in life, to be in the top is need "fight"...here is necessary make 4 games and then play against weak opponents. Great.

MrWCF: Wow!!!
I guess wow means what a great rating system this is?

Also, I can't seem to get the piece values in relation to each other straight--I'm still hung up on the 1-3-3.25-5-9 values in traditional chess.
I can't give you information of this kind with values etc. I can get my son speak about these things for hours, but i don't think that way in Chess and in Embassy Chess. It's just an experience matter and just think that way only.

I can give you countless general as also of specific type examples, of e.g Cardinal is better than Marshall here and there. And that of course with the small experience i got by playing this game here. In Chess is another matter and i have great experience being an IM. My beliefs about piece values-relations of this game have sounded a bit odd when i told them to another strong player, panzerschiff, but i think i'm right to all my conclusions. :-)
But this conversation is not for here but only to the appropriate discussion board i guess.

Pedro Martínez:Wow!!!

MrWCF: I've just found out that Jannssen's BKR was at 2100 already after the very first game he played. That was against Hrqls, who had played only 1 Cam game before that (he won against Robsaranga).

Constellation36:

BTW as per what Pedro Martinez says about your Embassy Chess skills(although seeing some of your games i saw some unwise moves<img>) do you mind playing some fast paced games with me?

This site is great but most top players are unwilling to play and
reject my invitations as they have to play some hundreds of games and
have no time for more. :-(

I would love to!!!
Additionally, I would really like to hear about my specific unwise moves--I'm still trying to learn more about Embassy Chess, especially about deciding whether castling is wise in this variant.  Also, I can't seem to get the piece values in relation to each other straight--I'm still hung up on the 1-3-3.25-5-9 values in traditional chess.  Walter Montego and I have had a couple of discussions related to castling and piece values in Embassy Chess, but I'm still hoping for more information.

Pedro Martínez:

MrWCF: How did you come to that conclusion? You said the player
in question had played four games against people whose ratings were
1341, 1094, "unrated" and "unrated". How do you know that those
"unrated" ratings were actually not very high ratings? I mean even if
you have an "unrated" rating, it is expressed in numbers in the BK
system. For example, you have finished, to date, 2 Embassy Chess games.
After a quick look at your profile, however, I can tell that your BKR
is somewhere between 2298 and 2445, yet it is shown as "unrated".

I know (or at least strongly surmise) that the unrated ratings were not high because the two games against unrated players were actually against the same player who happens to have a Cam record of one win and two losses.  That cannot possibly be a high provisional rating I would hope, right?  That fact, coupled with the ratings of the other two opponents, allowed me to question the 2100 provisional rating. 

Moreover, I don't really care about the 2100 rating, in itself; I am far more interested in how it was calculated.  I still don't have an answer to that question.

 

MrWCF: Let me hasten to add that this is not to cast aspersions on Jannssen's Cam skill level. For all I know, he is a world-class Cam player.

Is there any such thing as a world class player in Cam? Heh, i don't even know what Cam is, lol. :-)

BTW as per what Pedro Martinez says about your Embassy Chess skills(although seeing some of your games i saw some unwise moves) do you mind playing some fast paced games with me?
This site is great but most top players are unwilling to play and reject my invitations as they have to play some hundreds of games and have no time for more. :-(

MrWCF: How did you come to that conclusion? You said the player in question had played four games against people whose ratings were 1341, 1094, "unrated" and "unrated". How do you know that those "unrated" ratings were actually not very high ratings? I mean even if you have an "unrated" rating, it is expressed in numbers in the BK system. For example, you have finished, to date, 2 Embassy Chess games. After a quick look at your profile, however, I can tell that your BKR is somewhere between 2298 and 2445, yet it is shown as "unrated".

Thanks--I'm pretty familiar with the USCF formulae.  What I'm getting at is the fact that the use of the special USCF formula (for players with eight or fewer games) in this case would have resulted in a provisional rating equal to the average rating of the opponents in his four games plus 400.  That can't be 2100.  Some other mechanism must be being used on this site.

MrWCF: As far as I know, BK uses the following formulas:
http://math.bu.edu/people/mg/ratings/approx/approx.html

Fencer:Yes, I realize that provisional ratings are very inaccurate.  What I was asking is how they are calculated.  In the example that I posted, a player received a provisional rating of 2100 after four victories, zero defeats, and zero draws.  Those four victories were against players with ratings of 1341, 1094, and unrated (2 of them).  How was his resultant rating of 2100 calculated?

Undertaker.: For me, everybody must start each game rating with 1000 or 1200 points and each victory give between 10 and 50 points, in accordance with BKR opponent. For me would be the more correct rating calculation and would be necessary play many games and play against best players to be in the top.

IMO, that is really bad, and can cause really wild BKRs (like 3000+ BKR) easily. One should not only take into account the opponents BKR, but also how well estiblished that BKR is. If I play against someone who has 1500 more BKR points than me, but has only played a few games the last year, my BKR should not change much, regardless of the result. The point is that ratings should give an estimate of someones strength, but you can only estimate someone strength by comparing it to the strength of others. However, if you don't know someones strength very well (because someone hasn't played a lot), a result doesn't give much estimation of your strength, so your BKR should not change much.

In general, your BKR should change more the less established your BKR is, and the more your opponents BKR is. And it should change less the more established your BKR is, and the less established your opponents BKR is. And number of games finished is a poor estimation of how well established a BKR is. Number of recently finished games is a much better measurement than total number of games.

Ok, I'll say it one more time. Glicko.

rabbitoid: lol

wetware: prince? They must have upgraded the software. I still get the lawyer with the inheritance

Bernice: I have an acquaintance--a Nigerian prince!--who'd like to meet you.

"...he/she might be legit."

joshi tm: I think that isn't the point. For me, everybody must start each game rating with 1000 or 1200 points and each victory give between 10 and 50 points, in accordance with BKR opponent. For me would be the more correct rating calculation and would be necessary play many games and play against best players to be in the top.
Besides, it doesn't must be very difficult to do it (the new rating calculation, of course).

Fencer: I think 25 played games is still a bit low to predict a player's play rating. So why not make another rating level for played 100 rated games (ehm, well-established BKR)

Fencer: I know, but in this situation will be more easy for a player manage his rating, playing only with weak players and maintaining a high rating. On the other hand, a player with 2 defeats in first 6 games, for example, will have big difficulties to ascend quickly in same rating.

Undertaker.: Provisional ratings are VERY inaccurate and should not be taken too seriously. Established ratings should be always used to compare skills of players.

MrWCF: Welcome to BrainKing my friend.

I disagree with this rating calculation, because is very unfair. In Plakoto happened me the same thing. I did 4 games against players
with weak BKR and now i have 4 games, 4 victories and 2457 points. For me, a player with 4 games cannot be in first place.
This way is very easy to be in first places without big troubles. This isn't my idea about competition. For me, to be the first is need play many games and beat the best players and this don't happen here.

Constellation36: yeah con i can vouch for u with elizabeth she is a brilliant player but doesnt say a word not too sure about the others

Bernice: I didnt look too far but he has only been playing with people from Estonia.....there is a pisipaddy and a pisipaddy2...multi nic? seems a bit strange but he/she might be legit.

Let me tell you something. I'm playing with both elizabeth1932, pisipaddy, sema, NODDY and i can tell you that the playing style is the same, the time where these players are online/offline is the same, the behavior of these players is the same(they never talk a word).
Conclusions are yours.

Yet there is the slight possibility that they might be a couple of friends that come along and decided to play together here. But i doubt because of the playing style which is the exact same for all these 4 players.

And if you look at pmvaht's opponents where he obtained the 3000 rating in Backgammon, you will see that his opponents were the 4 players i have mentioned and been playing.

Can someone explain one aspect of ratings calculations to me?  How is an initial provisional rating calculated after four rated games have been completed? 

In the game of Cam, Jannssen has a provisional rating of 2100 after four victories, zero defeats, and zero draws.  Those four victories were against players with ratings of 1341, 1094, and unrated (2 of them).  How is a resultant rating of 2100 possible? 

Let me hasten to add that this is not to cast aspersions on Jannssen's Cam skill level.  For all I know, he is a world-class Cam player.

alanback: I didnt look too far but he has only been playing with people from Estonia.....there is a pisipaddy and a pisipaddy2...multi nic? seems a bit strange but he/she might be legit.

It's so impessive..until you take a moment to see how it came about.

Then, the only thing remarkable about it is that it's been allowed to stand.

SORRY I did not look at the set days off.

Hrqls: when the person doesnt time out but no vacation days are use .. then it might be that he has that day set as his weekend... or that a previous game has already timed out that same day and the day is already set as vacation. If people monitor the opponent's vacation days only just before the game between them times out, it may appear that the opponent didn't loose any vacation days when autovacation kicked in...

DeaD man WalkiN: when the person does timeout then he doesnt uses his vacation days automatically (its a setting wether you want it or not)

when the person is using automatic vacation days but still times out .. then he might be in a game/tournament which doesnt allow vacation days (red dot)

when the person doesnt time out but no vacation days are use .. then it might be that he has that day set as his weekend

Bluefin: found here:  Calendar

Bluefin: Settings / Calendar?