OpenChess

Alekhine defense challenge.
games:
https://pixeldrain.com/u/AjUXbNhQ

Thank you OrgZ !

ZamChess wrote: Sat Apr 05, 2025 7:25 am Thank you OrgZ !

Games:
https://pixeldrain.com/u/ND5btXqg

OrgZ wrote: Thu Jun 12, 2025 12:47 am

Code: Select all

Score of DarkSeid v6.1a SE_sse41 vs stockfish_25060220_x64_modern: 11 - 6 - 183 [0.512]
...      DarkSeid v6.1a SE_sse41 playing White: 9 - 1 - 90  [0.540] 100
...      DarkSeid v6.1a SE_sse41 playing Black: 2 - 5 - 93  [0.485] 100
...      White vs Black: 14 - 3 - 183  [0.527] 200
Elo difference: 8.7 +/- 14.0, LOS: 88.7 %, DrawRatio: 91.5 %
200 of 200 games finished.

Games:
https://pixeldrain.com/u/ND5btXqg

Nice...and Congrats!

Hello There,

Just out of curiosity, a new idea is born!

In short,
I could not stay far, and I realized to adapt
My current test to be close to Orgz recent test!

Actually I could not see much information about
Orgz used/played conditions..really limited info, e.g
In OrgZ played test: Only eng vers. and time control..
But what about rest? It is just a mystery...to predict!

In other words, it seems between SCCT vs OrgZ:
Same used cond. such as 2m+1s, same Eng vers.
But our machines speed diff. are not same, plus
We used different openings, mine is Unique etc.
Hashtable, Learning, Hyper-Threading etc may
Play as other role over results..sure if comparing!

But however,
Let's see what are the played 'results' under SCCT cond:

DarkSeid 6.1 wins again... Congratulations to T.R (OrgZ)!
But Elo difference is not so BIG as in recent OrgZ test!

But win is win !! moreover,
Very impressive that DarkSeid 6.1 is used older Evalfiles...
And despite, DarkSeid performs better..that means a lot!!
SCCT's Conditions / More Details:
2x Epyc 7B12, HT OFF, CuteChess, PonderOFF, 2m+1s, 1 Core, 128 Hash, 4 MEN
Both engines are used to play via Unique suite: each opening played twice
Plus DerkSeid 6.1 is played Eng Learning: OFF/ Read only on learning: ON

Above played GAMES soon as possible but after 2nd match,
Which will be with same used won openings by OrgZ's test
Note: I created a suite, based on 16 wins of Orgz test..
But 100% identical to OrgZ's used openings!

So later we will have some chances to compare the influences..
As we know, every eng performance depends on our used cond.

But at least I can say for righ now, after a quick checking :
OrgZ's used openings are not bad, but weaker, more critical
Than Unique openings, so the ERROR MARGIN can be too high..
Since if weaker lines are required many thousands of games
Per player..sure I refer for more 'right' Elo measurements!

Anyhow, I think also that my idea is not too bad..
Furthermore, it's worth giving a newer try, right ?)

And please stay tuned...more updates very soon ..

Greetings

UPDATE

1st of all,
What is going on? really unpredictable results via OrgZ suite,
The Elo difference is almost 22 Elo but in reality is it true?
Yes but it all depends, via serious strong lines, I hardly doubt..

On other hand, sure my fingers are already tired to explain,
But definitely there are experienced, who see the reality..
Such as tests via 'weak' lines, forget about to see the real
Elo strength, especially if playing small number of games!
For this I say, min. 4000-5000 games required (per player)
Even in case..why we have to spend so much time and efforts
For producing not so high-quality games? for my side not worth!
At least I say for most of my organizing SCCT tours..

Sure sometimes I run/use via weak..but not so often.. for example,
If always using in tests such critical lines, I can't sorry... because,
I'd prefer more draws, but as less games, e.g 1000 (per player),
In same time, definitely more correct Elo points on scene for sure!
Proved and guaranteed! right? sure any opinions, ideas welcomed,
But note also that according to experience over long past years,
Not only as now...in almost all tours, the stats say the 'last' words!

However, here are the latest played results (via OrgZ wins suite):
Where this time SF performed 22+ Elo better..Congrats to SF team!
Really great performance...I expected to see such unbalanced points,
But to be honest, not so so much..! anyhow..see 78% DRAWS !!

SCCT's Conditions / More Details:
2x Epyc 7B12, HT OFF, CuteChess, PonderOFF, 2m+1s, 1 Core, 128 Hash, 4 MEN
Both engines played via OrgZ wins suite, each opening played: four times
As before, DerkSeid 6.1 played Eng Learning OFF/ Read only on learning ON

GAMES:
https://mega.nz/file/rhonCZyS#zwztEOOzI ... 4snyekF5M8
NOTE: Download package includes exc. games plus the used openings as well

And I hope all my produced data to be helpful...but after all, please
Do not take too much seriously.. everyone is free, no one in chains!


Best,
Sedat