Hi. This message is for the owner of this website.

I intend to create a page similar to https://golem.tasuki.org/

To do so I'm planning to scrape the information from the HTML page   https://www.littlegolem.net/jsp/info/player_list.jsp

Of course with some system in place to minimize the load on the server (timed requests or something similar).

The person that I hired to look at the programming says that nevertheless scraping will put a load on the site and having some direct access or some type of API would be better.

Is that possible ? I'm agreeable to pay a small amount to allow for this to happen.

Why are people so obsessed with numbers?? Isn't this site for fun and about to know yourself? Why are people competing in their spare time against each other?? Isn't the economic system putting enough pressure on them?

So please don't load the server, instead invest your money into something actually valuable, like vacation or good food or a party with friends. Or just send it to Richard to support his great work for the community.

I like numbers, its how my head works

I love numbers too! And competition. And vacation. And good food & parties… Why can't we have it all?

I think competitiveness is part of human nature and “Agon” is one of the characteristic of games.

I would not say it is an obsession, but an healthy “push” to improve.

If anything, abstract games have taught me humbleness because no matter how good you are, there is always somebody / something better than you.

At any rate, I think it might be easier to contact Richard using the contact form at the very bottom of the page, he responds to that more often and that way things can be kept private as necessary.

Still, hiring someone to re-implement Monster Ratings? Quite a bit of enthusiasm you have there, it seems.

You could also try to arrange a third Monstership, of course. If you find enough players willing to invest the insane amount of time such a tournament requires…

Could be nice to have it!

Maybe it could even be an argument for new players to come here which is for me the most important point, get a bigger and bigger community of players.

I like numbers as well. Also it would be nice to have some other statistics here like the number of won championships and tournaments, players vs player stats, % of wins (as “white” and “black”) etc.

[q] Still, hiring someone to re-implement Monster Ratings? Quite a bit of enthusiasm you have there, it seems. [/q]

Maybe “hiring” was the wrong term. I simply set up a project on freelancer.com . I will pay a small amount of money for the project, but I have not technically “hired” and employee :-D

+1 passenger

Okay, this is your world, obviously. What remains for me to say is: Please, don't destroy my world by putting too much pressure on the server. As always: In doubt, don't!

Oh, and for all the number lovers, here is one as a gift for you guys: ({({},{})},{({({},{})},{})})

https://github.com/mrahn/lg

The project is ongoing, and we have a sample up and running, although with a few bugs to iron out. While doing this exercise, I noticed some “quirks” of the Little Golem Database.  The games are identified by a base name and a variant identifier i.e. Chess.Default.

I don't want to re-ignite the discussion on the ratings for Go; if Richard wants to keep the same ratings for all variants so be it. However the way the database is split seems strange:

GO.Size19x19, GO.Size37x37, GO.Toroidal Go, GO.Random100 and GO.Hahn are all variants of the game of GO.  Instead Go9x9.Size9x9 and Go13x13.Size13x13 are their own separate game.  So we have 3 “games” one with 5 variants and the other two with a single base variant each.  I think it would be better to have 1 “game” (i.e. GO) and 7 variants, especially as they all have a common rating.

Similarly HEX.Size11 and HEX.Size13 are variants of HEX, while Hex19.Size19 is its own game. We have 2 “games” , one with 2 variants and the other with  only the base variant. In this case, the ratings are the same for the two “variants” of HEX but different for the other “game” Hex19. I suggest it would be better to have only 1 “game” (i.e. HEX) and 3 variants. It is possible to keep the rating of the variants separate if desired, as illustrated by Draughts.

In fact Draughts.Dameo, Draughts.Checkers and Draughts.International are all variants of the game Draughts, but they keep separate ratings. This is the best solution in my opinion, with one “game” and 3 variants.

The idea I see behind having multiple “games” for different board sizes is to separate the games on the player page. For example, I am often interested to see my 9x9 Go games, but if they were all merged the games would get lost among my 19x19 games.

In my opinion, 9x9 can't reasonably be called just a variation, due to a larger strategic shift in the overall game. On the other hand it would be silly to have separate games (and maybe separate ratings!) for things like Go Random100 or Toroidal Go, which are notably less popular (and which most people don't play as one of their “main games”), but where skill still strongly correlates with regular Go skill.

In general I agree with Richard's design decisions, and would consider monster ratings a little off if they did not reflect what's on the site. Having a rating based on how LG works and a variation based on how you think it should work for Monster Ratings might be a good option. The default sort could be the latter option.

It is worth noting that WYPS variations have differing ratings as well. However, should they all be counted? Should only the English standard WYPS rating be counted? In my opinion, for Monster Ratings the best way to deal with variations like it and Draughts would be to only count the highest variation's rating, as that most accurately reflects what the player has invested their time in.

Just my two cents.

The site is up and running, a few bugs  not-wistanding (i.e the calculation of Go ratings is wrong). You can have a look at

http://dev.xportsoft-folio.com/home/index/2

Well, that certainly is something, it got done a lot faster than I thought it would! I like that it looks kinda similar to the old site, while still being new. Now if only I had some idea what the numbers actually mean! :)

Looks like a  sorting option using Monster Rating is missing too.

The sorting option for Monster Rating was added, as well as variations (which are password protected).

Now they need to correct the miscalculation of Go Ratings and to understand why some of the ratings are not downloaded.

Hi,

Quite impressive. Thanks for the work. I don't see any figures for some of the games: Einstein, Havannah, Catchup. will that be fixed in the future ?

Hi Bennok.

The missing games have been added, and the Go ratings corrected. The main page counts once each variant, which gives an high weight to some games. My favorite version is

http://dev.xportsoft-folio.com/home/index/2?var=No%20Luck%20No%20Word%20Games

Dots n Boxes please!:)

Will it be possible to create our own variant weighting systems? As a developer, I at least would be interested in doing something that better represents what I consider to be the actual skill of playing on LG.

Sean,  Dots n Boxes is included, you just need to scroll to the right.

mmKALLL, yes it will be possible, but all suggestion will have to be sent to me (by msg or on the forum) for preliminary screening.

Here the Link again, while I wait for permanent hosting.

Hi all, the permanent hosting is up and running.

www.golemrating.com

The main variation is ridiculous no?

It seems that games are counted multiple times.

Yes, I personally look only at

http://www.golemrating.com/home/index/2?var=No%20Luck%20No%20Word%20Games

But I am contemplating making the main page as the “weighted” variation.

Yes, it seems like having “weighted” as the main page variation would be a good idea.

It seems like in “weighted”, Go9x9 and 13x13 have a 100% weight, which possibly results in counting Go rating thrice?

As always, fantastic job on this! It looks really nice, love the domain name too. :)

Wow, nice! Also your monitor must be huuuge :)

I have double 32” monitors, but that's not the main reason for the layout :-)

One of the improvement in the works is actually to have a scroll-able table, while “freezing” the headers and the names.

Now my problem however is the cost of the hosting. For some reason Amazon virtual machine is much more expensive of what I thought (around 100 sgd for one month).

If you would be willing to host it back on the original server I can ask the developer to send the code.

Virtual machines might not deal well with the amount of requests necessary for something like this… It doesn't seem like an easy problem to optimize away.

A non-cloud service has its limits, sure, but that mostly translates to longer loading times rather than a huge bill. :)

Either way, 100 sgd/month sounds pretty normal for a VM used “in production” for a low-user database-request-heavy application like this. I would guess a non-cloud solution to still be around the 40-60 sgd range.

Improved version at:

www.golemrating.com

Now with scrollable table and indication of the weights.

Nice! Thanks!

Thanks!

http://www.golemrating.com

Are games that one doesn't have a score counted or not? It would be nice (if possible) to have a variation where not having played a game counts as 1500 (or less!)

http://www.golemrating.com

Ypercube, games without any played  should be counted as zero point, which is the same that you would get with a 1500 or less rating.

No negative points are awarded in any way (even below 1500)

So Maurizio, would it be possible to have a variation that score under 1500 is counted - and if a player has no score for a game, it counts as, say, 1200 (so -300)?

This variation would off course give advantage to players that play many (or all) games and hurt players that play only a few.

Website moved to:

http://golemrating.drfriendless.com/

This one should be more stable