Sandbagging
Sandbagging on a Go server (or in an over-the-board tournament) is deliberately setting your rank lower than it should be.
Jan De Wit wonders: what's the point of doing this? Surely it is not to increase your rating by winning lots of games; just setting your rating higher should do the trick. And if it is done only to avoid defeat, why not play at the Hippo's Room at Yahoo or some other low level server?
It makes me think of the motto of War Games with a twist: the only way not to lose is not to play at all :-)
HolIgor: Players sandbag for different reasons. Some like to win, some think that they are teaching and in some cases a pool of players that play mainly each other is formed and thus nobody makes progress. On IGS the number of double-digit kyu players is comparatively small. The median of the strength distribution is 2k. This means that half of the total pool of the players is at least that strong.
It is very naive to think that if you set your rank as 30k you'll quickly get to your true strength. It takes a lot of games (20 or more) to progress one level. You can estimate how many games a person has to play to progress by 10 levels. And if there are several runners like this they will make it difficult for each other. Still I believe that by the mere number of games people play on the Go servers their ratings are more accurate than in the real life, where you play two games per week in a club and sometimes in the tournaments. Mainly your rating is estimated by stronger players.
Grauniad: Relative ratings between players on Go servers may be more accurate than in real life for the reason you give, but there is a problem with the absolute rating because of the difficulty comparing with real life, as noted elsewhere on SL. (What am I saying: That the Internet is not real life? :-)
(moved here from "SandbaggersOnKGS" --Arno)
Scryer: Is it possible to sandbag on KGS, given that your rank is estimated by the server rather than by guessing it yourself? I started out losing some games because I wasn't used to the clock, and am gradually working my way up to what my strength was over the board 30 years ago. However, even if it's still showing me three or four stones worse than I think I once was, that may be a fair estimate of reality because I'm *still* not comfortable taking less time than I need to study a position, and I'm rusty to boot. Are KGS sandbaggers people who deliberately lose games to drop their rank? To what end?
tderz The term sandbagging assumes that the rating is lowered intentionally below the actual possible (non-internet) over-the-board strength. However, it is possible to lose a sequence of consecutive games without wanting it. As scryer above, I wasn't used to some unusual time settings and lost in my first 4 games 4 grades straight (wherefrom 3 on time). While, appearingly sandbaggering, I played every move à tempo (immedeately) in the last game against a 1dan, I found out that my played stones were still marked, yet my clock again already running, without that I could see my opponents moves. When these appeared some seconds later I would lose many seconds with each move. It was only this last game where I noticed it. Apparently this simply concerned a serious netlag as later confirmed by Bill Shubert. Hence, also netlagging can appear to cause sandbagging.
Cheyenne: Well.. there are people who keep their rank's low by either deliberately losing the game or by resigning. As for their motives, I couldn't really tell you.
One way to help identify a sandbagger (not saying this is a fully accurate method) is to look at their rank graph (on KGS it is part of their info). If the graph is fairly flat for a long time, then look at their games and see if they are alternating winning with losing games or if there are games that are resigned. As I said not a foolproof method, but it might help one at least know what they are getting into in a game.
And.. on the flip side (is there a flip side?) a request to sandbaggers.. please at the end of the game do a review and help the poor mark with their game.
I did have one good experience with a sandbagger (well maybe he wasn't a sandbagger).. a player of about equal rank (at that time about 17 kyu) joined a game I had up (I was white). After the opening my opponent proceeded to just slice me to pieces. I finally asked "okay -- how strong are you really?" at that point he resigned and then went into a review and then set up a teaching game. Turns out he had been playing for over 40 years. He said that he was a go gambler.
Blake: I disagree with the rank graph advice. My rank graph is fairly flat, but that's because I go for a long time without playing sometimes. Then there are the people who stall out at a certain rank and don't improve from there. And checking to see "if there are games that are resigned?" I'm not sure about that. Maybe if a majority of his games are resigned, I can see a problem, but when I'm losing, I usually resign. By your advice, I would be denied a game :P
- Assuming, of course, that you'd decline a sandbagger. I'd love to get sandbagged, but alas there are few who would sandbag against a KGS 1k (without handicaps, of course) :-( This would also get me games against stronger players without getting that annoying ~ after my rating..
Cheyenne: As I stated... it's not foolproof, and the flat graph basically is a flag to go check their game record
mgoetze: Well the only time I played a sandbagger, as far as I know, I just opened up a rated game (as a double-digit kyu) and i got an offer from a 1k at 6 stones. So I thought, hell why not. Well he totally took me apart, I resigned and he told me he was actually a KGS 4d. Wasn't really surprising to me because the speed at which he had read out the killing sequences was truly amazing (a 1k could probably have made the same moves, but not quite so fast). Well, no harm done. :) But I suppose I might have been a bit more upset if I had been expecting to have a fair chance to win the game.
Calvin: There are currently at least a few players in the double-digit kyu range on IGS who are as many as 10 stones stronger in other ranking systems and can perform well in even (and therefore unrated) games with players rated much higher on IGS, and I know for a fact that they are not all sandbaggers; they are either inconsistent players or play infrequently online. I point this out because you should not assume that a player is a sandbagger simply because you lose to him/her by a large margin, even though the relative rating suggests an even game. It could be that you could win if you accidentally hit upon the surprising big weakness that makes the player inconsistent, or it could be that the player is improving quickly offline and the rating system hasn't had a chance to gather enough data to adjust.
Vincent: One thing you can do to avoid accusations of sandbagging if you find you are stronger than your "offical" rank is to state your real strength in the game invitation.
mdh Ok, how do you keep from appearing to be sandbagging. For example. From the last US Go Congress I came out as a 14k. On IGS I can't seem to get past 18k. I sometimes pop up to 17 or 16. But mainly stay around 18k. On KGS I am currently a 10k. Soooo, for my next face to face tournament, do I enter closer to my KGS Rank or keep with the 14k. (You can't go lower then your rating in an AGA Tournament). 14k is about half way between my IGS and KGS Ratings.
Velobici: sometimes, I wonder if each entry shouldnt have a date...but that would make a mess of many pages. ;( As of 2 January 2006, mdh is rated 13k+ on IGS, and 13k? on KGS.
Neil: It seems to me that if you're playing in an AGA tournament you can't go wrong by using your AGA rating. It might be out of date, it might be wrong, but whatever it's flaws, the flaws are shared by all AGA ratings. Anyone who's used to the ratings and uses them will know those flaws and be able to take them into account. In other words, leave it to the AGA to worry about your rating's accuracy.
In discussions like this I find it most useful to remember the distinction between rank and rating.
Orando?: Well Kgs rankings aren't really all that great. I am sandbagging I guess, because I am about a 22/23 kyu player according to stronger players, but kgs says I am still a 30 kyu player. This is because you all start with 28 kyu and work your way up. So I took a 26 kyu player chalenge and I got 4 handicap stones, and I lost with 60,5 points. It was a good game, he was not a 26 kyu player, but my rank dropped deep because I lost big.
Tas: That is not how KGS ranks work. Your rank gets set after what wins and losses you have against your opponents "current" rank. That is if you play and lose, you rank goes down. If that player later turns out to be much stronger, the effect the loss has on your rank will diminish. Furthermore the amount of points doesn't matter, a win is a win, and a loss is a loss. Note that plays against [?] players does not effect your rank until that player himself gets a rank.
ggleblanc: I’ve been a double digit kyu for several years, and I’ve encountered many more sandbaggers on IGS than on KGS. What I have encountered on KGS is ranking variances based on time zone location.
Let’s assume that most people play Go when it’s convenient for them. For most people, that will be their local evening hours. With large enough groups in a particular ranking, subgroups of players will tend to play each other, defined by the time zones that they live in, and rarely play other subgroups.
When I’ve played at odd hours on weekends, I’ve noticed that the European double digit kyu players are four or five stones stronger than the American double digit kyu players with the same ranking. As mgoetze says, it can be annoying to play in a one-sided game when you were expecting a fair chance to win. Most of us only have a limited time to play.
Jef: I'm sandbagging. I guess. I used to have a KGS account and was severely overrated, and never found my actual rating before losing a reliable internet connection for a year and a half. Now I joined DGS and wrote in a rank lower than I know I am because I'd rather err on that side. Now I wish I was closer to what I think I am because there are other annoyances with being wrong on the low side. The nice thing is I'm playing a fella and telling him how to (maybe) beat me in a handicap game, which I know he's not used to.
Mef: I wouldn't call this sandbagging, because it's not something you are doing intentionally. Guessing low for a rank, but then playing honestly until you stabilize at a real rank is fine, especially since most servers will adjust how your old games are scored based on your current rank, so once you have settled there's no real harm.
Mihoda: I know of at least one case on KGS where a player has established a rank 3-4 stones weaker than his playing strength, and then proceeded to play every game free. Playing games free means that his rank is kept intentionally worse than his playing strength. Presumably he is doing this simply so he can win games with grossly unfair handicaps.
Tapir: "I’ve encountered many more sandbaggers on IGS than on KGS." - the main fault of the IGS system is that you're sandbagging, if you don't play a large number of games or you do play offline or on other servers as well while improving. As KGS is much more flexible in adjusting your rank the ranks there are less underestimated. - Once in a while I still use my IGS account (9 stones weaker than my KGS account) playing in the spirit of a teaching game trying to make the gap smaller. A re-evaluate rank function for players with >70-80% winning ratio might be helpful against such unintended sandbagging but otherwise you have to play the games - and frustrate weaker players by this.
robw: The other day my friend started up a new account and played a guy that was listed as 27k. He destroyed my friend and then said that he is really 2D, but he was playing on his friends account. I guess some people out there are just boosting their buddies scores (lame).
Related Links: