A decision has been made on the official map generator

[Janzert]

Since this last set of maps (map_r538_*) has been in place for about as long as the finals are likely to last I've now switched to a new set of 75 maps. These maps were generated using the same generator as last time without any changes.

The changeover to the new set occurred at 2010-11-22 13:03 with .

Janzert

[Accoun]

[antimatroid]

There is a trade off to having too many maps, if you do that, very few games get played on each map, you get more consistent results if lots of bots play on the same map. You also don't want too few maps because then whichever bot happened to be best on the couple that were generated wins. This has been discussed a number of times, I expect Janzert will manage a fairly nice mix in the middle for the maps generated at the end.

Also, with respect to rankings, while I agree that to get accurate rankings on these maps one would need that, it's not the final contest yet so they're not really that important. The rankings will be reset before the final tournament is run.

[antimatroid]

A 999 map set would give the same winner if all bots played each other on all maps, but that simply isn't feasible. If you have 999 maps, the top 10 bots are probably quite unlikely to play each other all on the same map, so basically the rankings based off those games are a bit of luck as to who got paired against who on which map, that leads to similar random problems with rankings. The ideal number of maps is somewhere in the middle, having too few / non random maps would be the worst, but having too many is also an issue.

As for Janzert, I think you would find he's a really nice guy, has put a lot of his own time into ensuring the contest runs as smoothly as possible while fixing up quite a few things himself. If you really don't like the competition so much, why are you competing?

[zandee]

Hi,

I'm aware of a number of users (including myself) that were unaware that the forums contained posts like these, meaning that bots have been tweaked to the old map generation rules.

Notifications like this would be better sent to all contestants out of fairness via a mass email. This has left those users having read this post with an unfair advantage as all the demonstration maps were a lot smaller than the ones I have seen played in the tournament so far, and therefore require an entirely different strategy.

Finding this post at this stage has infuriated me and as a result I have lost all confidence in the organisers.

[Janzert]

These maps have been in use on the main server for the past 2 weeks, surely if it makes such a difference for your bot you noticed the maps it was playing on?

Janzert

[zandee]

Some of the maps are noticeably different from the ones played over the last few weeks (and yes, I had been paying attention). In the majority of them, the planets are more numerous, and the growth rates are much less evenly distributed. In one such map I just saw, nearly every planet had a growth rate of 5. This map did not exist in the previous allocation.

I'm sure you can do a simple database query to find a list of maps that I had not played before submissions closed.

[antimatroid]

The map generator used to make these is exactly the same as the one used to make the maps used on the server for the last two weeks, I assure you maps have had up to 30 planets on them with line symmetry that whole time. Also, go recheck the specifications, it never said anywhere that the old ones are the only type of maps to be used. People keep knocking around the organisers when they have done nothing wrong.

[glevee]

[aerique]