Life In 19x19

Tygem has probably the best available ranking system. It will take something like 20 straight lost games to fall to 3d, another 18-20 for every other two ranks (considerably longer if one could win at least 5 games, but we are talking about situations when win has more like 0,01% of probability).

Gorim wrote:Tygem has probably the best available ranking system. It will take something like 20 straight lost games to fall to 3d, another 18-20 for every other two ranks (considerably longer if one could win at least 5 games, but we are talking about situations when win has more like 0,01% of probability).

That doesn't sound like the best available ranking system to me. Many rating systems have a notion of rating confidence: if a player doesn't have many recent games, their rank shifts more quickly. If a player claims to be 5d but has no wins at 5d, they could drop to 4d in 2 or 3 games.

emeraldemon wrote:

Gorim wrote:Tygem has probably the best available ranking system. It will take something like 20 straight lost games to fall to 3d, another 18-20 for every other two ranks (considerably longer if one could win at least 5 games, but we are talking about situations when win has more like 0,01% of probability).

That doesn't sound like the best available ranking system to me. Many rating systems have a notion of rating confidence: if a player doesn't have many recent games, their rank shifts more quickly. If a player claims to be 5d but has no wins at 5d, they could drop to 4d in 2 or 3 games.

Agreed. To me, the Tygem rating system is one of the worst possible systems, given that people can specify their initial rank.

But what some people value above all else in rating systems is transparency, and Tygem has that.

I've started to really appreciate Tygem's system. I probably prefer IGS' in some ways, although I'd make a few modifications to increase the granularity of the increments. I prefer both now to KGS' system.

Let me ease your mind by telling you that I have stopped doing this

In the mean time I have more than 20+ lost games on my track as 3d on tygem and none won
Maybe I will quickly share my experience in those 20 games.

So my goal in these games is usualy to keep alive while trying to get more than i should aka by fighting. Of course they can read better. I ended up resigning pretty quickly and then reviewed my moves and tried to understand why my positions got compromised. I analysed my shapes (most likely this was the problem) and looked for reading mistakes. What I noticed though is that it felt to me like a 5d does not have that much of better complete strategy. (although this probably is a serious illusion ) It seemed to me like what is the difference between me and them was shape and reading.

In the mean time I have also played a buch of 3k's and a 1d on kgs which offered me a teaching game which was fun. I think in time I can learn the same things I learned from playing reverse sandbag games on tygem by playing teaching games on kgs.

I'll miss it but I know its not the right thing to do.

Cheers,
Otenki

My impression is that most Tygem ~4d players aren't very good at strategy.

3 wins and 2 losses as a 5dan for this kyu player. Plus one extra win because my opponent got bored for me playing so slowly. Yeah it seems I am the one doing all the sandbagging.