3 major blitz sites--observations

Originally posted by no1marauder
It's interesting; I've played 50 some odd timed games (30-90 minutes) on both playchess.com and FICS and my rating at playchess is about a 100 points higher (high 1800's to high 1700's). I've never seen any good evidence of computer abuse on either in my analysis of my games. I don't have that many games on USCL (less than 20 and I'm rating low 1800's) b ...[text shortened]... 00's at all three sites is that the USCL ratings are probably 100-200 points overrated.

I've played fewer than 100 "standard" games at each of playchess (rating mid-1600s after a string of lossses) and FICS (low 1700s), but have played more than 300 such games on USCL (mid-1800s, with a high of 1951 when I was in the top 100 of that rating there). At USCL I've played twice as many blitz games as standard (blitz rating high-1800s), but have played over 1000 at playchess (rating peaked at 1814 two days ago, currently high 1700s), and over 5000 at FICS (rating hovers near 1600, also my USCF OTB rating).

I don't have clear evidence of cheating on FICS, and do not subject my suspicions to the consistent, objective testing that is your habit. This is especially true where I mostly play 3 0.

Two years ago I read an article published by some computer scientists who had successfully hacked the time-stamp system at ICC. Their article recommended a number of additional safeguards to reduce the possibility of cheating by these means. I suspect that FICS has been less diligent in implementation of these protections.

With alarming regularity on FICS, I lose on time from an advantage in material and/or position as well as on the clock when my opponent starts lagging. These losses could be causes by some form of time cheating akin to that described by the white hat hackers mentioned above, or they could be caused by breaks in my rhythm due to the slowing of the game. I often get the sense that such lag could be caused by engine use, but even a weak time-stamp system should prevent CPU memory issues from affecting the game clock. Of course, sometimes it takes longer to find the right moves in a winning position than to move in a lost position.

I know that playchess and ICC are particularly aggressive in pursuing cheaters, while FICS is run by an all volunteer force. They investigate cheating, but I'm not sure how aggressively. USCL simply has too few users for any valid generalizations.

The relatively lower ratings on FICS could be a consequence of a weaker pool to play against. There are few GMs, IMs, FMs, and NMs on FICS, and fewer non-titled players that play blitz as well as the titled players. ICC, playchess, and even USCL all have titled players in abundance.

Comparison across rating pools is always fraught with difficulty. For server-based, correspondence-paced chess, I find my lowest ratings at RHP. Here and GK are my only established server ratings lower than my IECC (1750) and USCF (1805) Correspondence ratings.

Originally posted by Oddjob291
Just wondering, do these blitz/bullet games actually benefit you besides calculating quickly.

I'm not sure they even help with quick calculation. I much prefer long slow games, but circumstances don't usually allow me to commit a couple hours to playing without the fear of being pulled away by kids, business or high maintanace spouse.

PS~ The original message should have stated a standard rating of 1700 vs. a blitz rating in the mid 1100's.

Originally posted by buddy2
The other thing i've noted is that icc and playchess players stick to recognized lines, follow opening variations quite far. On FICS openings don't seem to be as important as tactics, see alot of double fianchettoes and homemade stuff. Actually, this avoidance of theory makes them much harder to beat.

just grab the center and start taking the board one square at a time. The trick is to develop the ability to hammer out more than 60 reasonable moves in three minutes.