The solution isnt bump everyone’s rank up, its getting stronger players to play more on OGS. I’ve heard often that “once you get to OGS 3k-1d leave and play on tygem/fox”. There are a lot of botters at the mid to high dan ranks and some new accounts are highly suspicious. The ranks themselves are fine.
1 Like
I tried to download Fox once, out of curiosity because people said that players there play very aggressively and I wanted to see what’s up and the antivirus flagged it as “dangerous” and sent it into quarantine.
If you go to tygem’s website, it is not even https and the browser says “unsafe connection”… what’s up with that? I haven’t seen a website without https since… well it has been so long that I cannot even remember.
2 Likes
I dont know off the top of my head, but there are people who have made versions like minifox and weiqihub that can let you access fox.
3 Likes
2d egf here, and 2d ogs as well. Most dan players I know IRL (and I know a lot of them, including their online handle) have the same rank, be it here or IRL.
I don’t see the point in reevaluating at the moment.
The lower dan ranks on fox are overranked compared to EGF, by a lot, that’s a fact indeed.
As a 2d EGF, I can give Fox 2d anywhere between 6 and 9 stones. Do we really want to devaluate western ranks ? Probably not.
5 Likes
Assuming OGS rank = EGF rank for dan players, then OGS ranks need a bit readjustement since 1d OGS was designed to be 0.7 stone weaker than 1d EGF:
I can’t confirm your observations, dan players I know generally don’t have recent rated games on OGS, and the few who do have a lower rank on OGS than EGF.
2 Likes
It was almost a year ago when I downloaded game data and crunched some numbers…
I have to dig up the code from my old PC, clean it up to open source it and maybe give the statistics stuff another try.
Note: Deflation, if there is any, might be different across all the different ranks.
2 Likes
Let’s keep this quote and share it every time this topic comes up 
2 Likes
The thing is :
I am 2d EGF, sometimes I play very well. Sometimes I don’t. I can beat a 4d if they underperform (and I perform well), just like I can lose to a 1k if they are on a good day and I am not.
The 0,7 difference is the ideal difference, it doesn’t mean it should or will always be 0,7.
It can be 0,2 on some days, just like it can be 1,2.
So yeah, our strength is not always the same. I can be 1,5d one day, and 2,5d the next one.
If my opponents is 3,3 dan for example, he can very well play at the range of 2,5- 4,5d
Just like I can play in the range of 1 to 3d. If the conditions are met, then yes, the games will be balanced. That’s how the ranking works. If this wasn’t the case, you would never be able to beat a stronger players.
Imagine it like this
1 Like
I’m aware that ranks fluctuate, so the 0.7 stone difference should be understood as the average difference. If this were accurate, then among players around 1d, there should be more of them with a higher rank on OGS than their EGF rank than vice-versa. My observations tend to show the opposite, but perhaps my sample is too small. I’ve added players to my table, all of them have played a significant number of rated games recently on OGS.
| EGF rank | OGS rank |
|---|---|
| 1k | 2k |
| 1k | 1k |
| 1d | 2k |
| 2d | 2k |
| 1k | 2k |
| 1k | 2k |
| 1k | 2k |
| 1d | 2k |
| 2k | 1d |
| 4d | 1d |
| 4d | 2d |
| 3d | 2d |
| 3d | 2d |
| 2k | 1d |
| 1d | 1d |
| 2d | 1d |
| 4d | 2d |
| 2d | 2d |
| 4d | 2d |
3 Likes
And these are mostly French players I presume, who generally are quite strong for their EGF rank IME (not played in real life since 2019).
2 Likes
mostly french yes, that’s where I play in tournaments the most, but I also know quite a bunch from other countries
Yes, all but three are French, however I don’t think that French players are stronger than their EGF rank anymore since the FFG rating system update a few years ago. I checked the last EGC main tournament, winrate was
France: 53%
UK: 47%
Germany: 48%
so the difference doesn’t seem significant.
2 Likes
Yes, my experience is quite out-of-date, e.g. a French 3k at the London Open in 2010 might be a British 1d.
1 Like
Is that a mere guess? Would be interesting to see some maths to judge the significance.
I said “doesn’t seem” because I didn’t make any precise mathematical analysis. The number of games played by French players was 417. If we compare with tossing a fair coin 417 times, the standard deviation would be about 2.5%, so a winrate of 53% doesn’t seem statistically significant.
But some matches were French vs. French, and I didn’t count how many, so a finer analysis would be needed. So I just gave my feeling.
2 Likes
A crazy idea is to just add a button in the settings wich switches to „Japanese ranks“. If activated all displayed ranks are shown around two ranks higher.
I think Pandanet does this.
The confusion about ranks not being the same across devices would push confused forum posts by around 5%.
But players, which prefer the less strict system would be happy 
8 Likes
You are so evil, godeon 
1 Like
Besides, beyond the technical question of statistical significance (= “Is the French win rate of 53% their true, predictable winning rate we should expect in future games, or is it spurious”), it’s also meaningless, since a 53% win frequency is just a few Elo points difference even if it’s real.
I agree with you. I have nothing to add to this. In the end, it’s a decision for the managers here to make about what to do with the levels here.
I agree with you that we Danes have bad days, we can play like Kew. Personally, I don’t play at my full strength in every game, but I have a streak of days. I can lose completely and I can win completely, depending on the day and season.