rating system

While I just love the game – and don’t care so much about ratings (on the msn chess site – we all lost our rating records a couple of weeks ago – and I still play there with friends and those I meet there). I guess the trick with the Glicko rating system used here would be to practice up on other sites … wait for your RD to go up … and then when you win here – via the Glicko rating system’s short-sited assumption – your rating will rocket up.

What I mean – is that the Glicko system factors in time since your last appearance on a given system that uses their calculated records – and assumes that you have not played since … and assumes that you are coming in cold - and therefore gives you a calculated 'handicap' on that basis. Instead – you could be coming into a Glicko ratings system chess site hottt – from another site, …, – and be more prepared than the other player who only plays on that given site.

While Dr. Glickman makes sense from a non-real-world perspective – assuming that you can and only will play on that given system – the only way to really be able to factor time into such calculations with any degree of certainty of reliable assessments as to the true nature of the two competitors in the pre-match condition (and then to reward and penalize them both on that basis depending on the outcome) – is to lock up both their computers, their chess boards (, and pads of paper in case they might decide to play the short-hand way) – until it is time to come online to play on that site which uses the Glicko system – and that would seem a little too ‘totalitarianistic' to me :)

Anyway – being an engineer – and knowing that usually the biggest source of errors made is during the time one models and makes assumptions of the physical real-world-situation - before you start to crunch the numbers. I thought it appropriate to suggest that his initial assumptions fall short from the real world - even when considering those - who with no intent to defraud – will play other people and friends on other sites – or in person – as I still will end up doing – even after I start to play here.

Anyway – looking forward to having fun playing chess here – some interesting features here not available on the msn chess site – although – the games would seem to be at a faster pace there – as they assumed ;) that players who play there – would sooner just play out a game in one sitting. BTW – I have yet to play a game here – was invited here by a chess playing friend on msn – so – was just browsing around this site – when I caught some of these comments – and thought I’d contribute.

And for those who are still asking why their ratings did this or that – (with the exception of Gary Seven’s situation – wouldn’t like if that happened either – could be someone within the system – as on the msn site – I had one guy of a MUCH higher rating beat cold – he then commented that he will get the points anyway - and proceeded to let his remaining 10+ minutes run out – and he did get the points) – check out the link that Erik provides that shows up as,

"

The Glicko System by Professor Mark Glickman, Boston University", and look for RD – and read up on that part.

Erik – well put – (and Dr. Fallet – good points of contention – here is one more :)

Thanks, Brent, B.Sc.M.E.

Like you I don't put any stock in the whole ratings thing, its for the professionals, probably works fine within those parameters. To me its nothing more than a current form guide.  I'm sure there's a guy on this website with a rating of 3000+ which is higher than Kasparov so its not realistic. All that tells me is that this guy knows how to play and is used to winning.

Agreed DMX21x1 - and not only knows how to play - but knows how to play the system ;)  For me - i just love the game - and do relatively well - on average :)

The Glicko system assumes nothing of the sort. It is a measure of uncertainty, but it does not assume a direction in that uncertainty. The system allows for you either getting rusty from not playing or getting better from playing outside the system.

In fact, if you have been playing outside the system and have been getting better, then your rating going up quickly is what you want. That way it will more accurately represent your chances of winning within the system.

And no, DMX21x1, comparing apples to oranges will not give you realistic results.