So I was just wondering who is the player on lichess who has beated big guys while having a relatively low rating, a GM killer hopefully.

I came up with this ratio in order to try to figure it out:

Pick up your top 5 victories against high-ranked opponents (I personally focus more on rapid, but you can do it with any time control). After that add all of their ratings up and then divide it by 5.

Now divide that number by your highest ever ranking in that particular time control. A higher rating of big guys beaten will end up in a higher ratio, and a lower own highest ranking will end up in also a higher ratio, thus regarding players who have beaten high ranked opponents while maintaining low rankings themselves as the big guy killers.

An example:

I'll pick rapid time control for this.

my best 5 rated victories are : 2078, 2025, 1948, 1932, 1924.

So I'll average those: (2078+2025+1948+1932+1924)/5= 1981.4

My highest ever ranking in rapid is 1694.

So the ratio is 1981.4/1694 = 1.1697

So there it is! Let's find who's the big guy killer!

Ps: Specify time control please.

Thanks.

It seems my ratio rewards lower ranked players much more.

As when you get higher up, beating someone 200 rating points above you is much more difficult than when you're 1200.

Or it seems to me it is that way.

1.00009 Rapid (2114/2029)

@daniel921126

What if you got the average rating for the top five victories, and then converted that and your highest rating into the corresponding percentiles and then subtract those percentiles to get a percentile difference between you and the average of your top 5 victories?

Example: 1994 is about the 92nd percentile in antichess, and is my highest antichess rating. My top five victories in antichess have an average rating of 2324.6, which corresponds (approximately) to the top 1% in lichess antichess (99th percentile), resulting in a percent difference of 7%.

By the way, using your formula (for antichess) I get 1.1568.

Also, most antichess games I play are blitz or superblitz time control.

Nice idea, by the way!

@the_loving_boy thanks for your ratio

@Jacob531 it's also a very nice idea.

Let me do mine quickly:

My highest rating in rapid of 1694 has a percentile of 75%.

My average top five victories have a rating of 1981.4 which is a percentile of approximately 95%.

Therefore my percentile difference with the strongest guys I have ever beaten is of 20%.

Also I would like to hear your ideas about the ratio.

I think it should blow up in slower time controls. Also, that it rewards aggresive play. You might lose against a 1200, but if you beat a 1900 a couple of times, your ratio should go up.

I don't like the original ratio for the reason you mentioned, but I'm not so sure that I agree with you about style and time controls. I'm not sure style would matter, and I would definitely think that upsets would be more common in faster chess.

Just my 2 cents.

@daniel921126

thanks for your insights!