Talk:RoboRumble/Archive20130317

From Robowiki
< Talk:RoboRumble
Revision as of 09:12, 17 March 2013 by Chase-san (talk | contribs) (archiving)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Subpages

Should we replace ALL of the subpages? or just the 'important' ones? -- Starrynte 22:16, 17 November 2007 (UTC)

Name

Since the new wiki supports symbols in page names, should we put this page at RoboRumble@Home instead of RoboRumble? We already have a redirect from that page to this one... --AaronR 23:34, 17 November 2007 (UTC)

I second this change. » Nat | Talk » 02:30, 28 April 2009 (UTC)

RR seems to be down

Roborumble is returning 503 Unavailable for all pages today. Not sure why. -- Synapse 03:07, 19 September 2008 (UTC)

Are you sure? Both rumble.fervir.com and ABC's work for me. -- KetsuNfwu 03:33, 19 September 2008 (UTC)
I fail. I meant Robocode Repository. -- Synapse 05:44, 19 September 2008 (UTC)
Not a big surprise there. =P -- KetsuNfwu 13:21, 19 September 2008 (UTC)

Hi! What the heck happened to RR@H again? The main roborumble has only ~60 bots again... Otherwise, I uploaded Pugio 1.2 1 week ago and has only 120 battles in nanoRumble. Situation is the same in Darcanuck's page, too. Nobody is contributing?--HUNRobar 12:30, 11 October 2008 (UTC)

Both ABC's server and DarkCanuck's server have no full pairings yet. Once the megabots have full pairings, the minirumble will get filled completely and as last the micro- and nano-rumble. Be patient, but it probably requires a few weeks before all nanobots have 1000 battles. Remember, 640 * 640 bots means a lot of battles! Pulsar's server has another issue. I don't know what happened, but I still run (repair)battles there. --GrubbmGait 13:14, 11 October 2008 (UTC)

I realised that in Darcanuck's server the older versions of bots are still in game. I think the competitors there should be cleaned up. For example, I don't want to generate battles for Pugio 1.0 when I wait for the results of 1.2. --HUNRobar 14:57, 11 October 2008 (UTC)

That's a temporary thing -- I'm uploading a huge amount of data from ABC's server which includes old participants. There's a big performance hit to activating/retiring participants repeatedly, so for now everyone's in the rankings until I get caught up. But don't worry, no rumble clients will actually run these bots. They'll use only the version from the participants list. In a few days you should see the old bots disappear from the rankings. --Darkcanuck 15:03, 11 October 2008 (UTC)

Whoops edit conflict, accidentally overwrote Darkcanuck's reply with this: "What versions get battles doesn't depend on the server, it depends on the participants list that clients are configured to (which should be the old wiki's one, the one here has not been synced for a long time). The server would have nothing to do with that so you shouldn't be generating battles for Pugio 1.0 unless your client is misconfigured I believe." Anyways Darkcanuck's reply clearly explains why there are currently old versions listed :) --Rednaxela 15:10, 11 October 2008 (UTC)

Ok, thanks but I'm very excited about the results of Pugio 1.2. :)--HUNRobar 17:40, 11 October 2008 (UTC)

Why can't I access ABC's server? My connection get timeout before it finished loading! -Nat 12:42, 6 January 2009 (UTC)

Lynch the Multi-Threaders!

A thought just occurred to me. I have never liked the fact that robots could be threaded. Why?

  1. On a single-core (non-hyperthreaded) machine it cannot help the bot without hurting the opponent. Either a) all computation is done on one CPU during your turn, in which case you could have done the same computation with less overhead without the threads, or b) you do computation past your turn, eating up CPU cycles from your opponent.
  2. On a multi-core (or hyperthreaded) machine, it violates the idea that all bots are equal except for their AI. You suddenly get 2 CPU's in your robot, essentially.

This is a reality we have had to live with, because that's the way robocode is designed. HOWEVER - it just occurred to me that we could enforce a rule saying that no multi-threaded bot is allowed in the rumble! Which I hereby propose. Besides being unfair to the bot's opponent, acting like a virus stealing CPU, it steals CPU from bots in OTHER battles now-a-days when more than one rumble client executes at a time on multi-core machines! Any bot caught using multiple threads would be immediately removed from the rumble. Does anybody agree? --Simonton 02:47, 23 September 2008 (UTC)

Agreed. It isn't fair on the opposition, and makes it impossible to design a bot within the CPU-constant constraints. --Skilgannon 10:34, 23 September 2008 (UTC)

I either didn't see or forgot about this section. There are already bots that use multiple threads in the rumble: Toad definitely does, possibly X2 (which shares some Toad code), and probably others. Not sure how I feel about removing them at this point, just contributing some facts here... --Voidious 02:20, 7 August 2009 (UTC)

Well, there are java APIs that allow things like only measuring the time used in a particular thread. Robocode could start to use these in order to make this fair, AND as a major bonus it would make using Robocode on a computer with a heavily loaded CPU not cause skipped turns. The issue is... those APIs don't give the same kind of resolution, thus meaning that the only way we do it would be if we calculated something more like "How much CPU time did threads of this bot use in the last 20 ticks?", instead of "How much time did this robot take between the start and end of this turn?" which would make skipped turns behave a bit different. Personally I think that may not be such a bad thing, but I'm not really entirely sure... --Rednaxela 02:51, 7 August 2009 (UTC)

It looks like this » Nat | Talk » 12:53, 7 August 2009 (UTC)
Hm? "ERROR This artifact is not accessible." is what that page gives --Rednaxela 13:11, 7 August 2009 (UTC)
Sorry, I never noticed that this is 'private' tracker... Fnl thought that he needs more precise time and mem.usage detection. I have a comment on some time measuring much like to your "How much CPU time did threads of this bot use in the last 20 ticks?". However, we didn't mention a thread at all. But I believe that Robocode allow Robots to create up to 5 threads, I'll email Fnl this. » Nat | Talk » 13:53, 7 August 2009 (UTC)
The tracker is public now. :-) Basically, I want to measure the CPU usage with the new features in Java 5.0 for measuring the exact CPU per thread. Hence we will (when using the TheadMXBean) know for a fact exactly how much time all threads in a robot is using in total. However, the time interval when measuring using the ThreadMXBean is big compared to what we are used to with the CPU constant currently used in Robocode. So we need a solution that can deal with this issue. I have waited with this improved feature as this will have a big impact on the scores for the robots in the RoboRumble. We have already seen how much difference the getNextVelocity() makes, so imagine the difference when we make change the concept behind the CPU constant. --Fnl 21:07, 7 August 2009 (UTC)
Personally, I'm actually not too worried about the impact of this change. Most bots don't skip many turns at all, and the effects of a few skipped turns are often overstated. As long as the average amount of CPU allowed per tick is still similar, I can't imagine it affecting scores too much. I suppose it might have a big impact on bots that abuse the currently untracked multi-threading, but I think we can all agree that they're sort of abusing the system right now, anyway. Of course, I'd be happy to run extensive tests, too. =) --Voidious 21:55, 7 August 2009 (UTC)
I'm agreement with Voidious on this. I doubt there would be any problems caused by this so long as the correct amount of time was chosen. I would say though, that I think that this feature should be targeted at a 1.7.1.5 release, for the same reason that it's generally considered bad practice to change both a bot's movement and gun in the same rumble iteration, haha. As a side note, I wonder if while we improve the measurement method to not be affected by other things on the system, that maybe we should also improve the cpu constant calculation. For one, it should use the same ThreadMXBean methods to be accurate, but also I wonder if we could have a more realistic benchmark of what a typical bot actually does with it's time. I propose we include a small suite of some of the things a more cpu intensive bot does in order to create a more representative benchmark. I propose coding a small suite consisting of something like the following: 1) KD-Tree, 2) Nanobot-style pattern matching, 3) Array operations (searching for peak, addition of a gaussian curve to an array, multidimensional array access), 4) 'Play-It-Forward' simulation, 5) GF calculation (aka simple trig). I think I'll begin a page for comments on and code for an improved benchmark --Rednaxela 00:00, 8 August 2009 (UTC)
1.7.1.5? It was there for about eight months. I think maybe 1.8 or so... » Nat | Talk » 06:10, 8 August 2009 (UTC)

What Now?

How can I submit the new versions of DrussGT and Cunobelin? Has anybody heard any news on the old wiki? Will it be recovered? Is it gone forever? --Skilgannon 15:47, 8 December 2008 (UTC)

Well on November 30th I emailed PEZ and got a reply that it's an ugly error and that he'd look into it. I haven't heard anything since but I guess I might email PEZ again. I certainly hope it won't be gone forever as I'd at least like to have a read-only version due to how much great stuff there is there that's unmigrated. --Rednaxela 16:12, 8 December 2008 (UTC)

I really hope the old wiki content still lives on. In the meantime we could agree to use this wiki's list for my server. I don't have time this week to make the proposed participant list changes, but I could lock out add/removes for all but a few select clients to make sure we all use the same list? --Darkcanuck 18:28, 8 December 2008 (UTC)

Well, the old wiki is back! I do wonder though, should we switch to using this wiki's participants list anyway? or should that wait till some other day in the future? --Rednaxela 18:17, 9 December 2008 (UTC)

I'd like to keep using the old wiki at least until I can code up something to validate participants on the server side. Then it would be easier to switch and not have to worry about some clients using the wrong list. --Darkcanuck 05:39, 10 December 2008 (UTC)

Random question: How much do you think the local client would speed up if I compiled it natively? Right now, we haev a bit of a bottleneck figuring out battles andif we could speed it up by a factor or 3 or so, why not? --Miked0801 21:52, 25 May 2009 (UTC)

Nanos get little love when things get busy

I've been waiting for nearly 4 days for my latest bot to get to 2k battles against nanos, and I'm probably gonna have to wait 3 more for it to happen. Why? Nanos, after they get to their 1 battle per bot limit, are the last bots to be updated under the current system. I understand why the system was done this way, but is there anyway we can add a time component to the battle chooser to make sure that when things heat up in the rumble, nanos get some love as well? --Miked0801 20:47, 19 June 2009 (UTC)

I hear ya. Even in Mega-land, it can take a while to get to 2k battles, and I don't really trust a rating until it does (even with all pairings). There are some RR client options you could try, like excluding bots or running NanoRumble only, but the latter will just do random battles, I think. A couple of nice client side options, imo, would be mini/micro/nano while still honoring pairing priority, and ignoring pairings while running bots with under NUMBATTLES (in all applicable rumbles). I'm definitely in favor of RR contributors having control over what battles they want to run... --Voidious 22:48, 19 June 2009 (UTC)

Nanos get way more than their fair share of processing time. By a time a nano reaches 2000 battles in its own class, it has typically reached 5000+ battles in the main rumble. Few, if any, megabots have yet to reach that level. Not only is it unfair to folks releasing larger bots, it also takes away cpu time from other nanos. So should we abolish nanos? Of course not!

I'd vote for changing the client's battle selection such that when a new bot reaches full pairings, the remainder of its 2000 battles are fought against opponents in its own size class. Those battles will still count for the higher classes, so a nano would really only need 2000 battles overall. This would mean faster results for nanos and more cpu time for everyone else, regardless of size. --Darkcanuck 03:19, 20 June 2009 (UTC)

I second Darkcanuck's suggestion, as it seems most fair to me. :) --Rednaxela 05:12, 20 June 2009 (UTC)

And yet, some nanos love to beat up on their larger sibling... But yes, when things slow down, nanos do get a ton of extra battles, but to be fair, most nanos run much, much faster than megas so the hit isn't that bad. My main concern was waiting at the back of a line of never ending micro/mega selections. Impatients and such. If we're going to change things to minimize waits, then first completing pairings, then 2k in class, then low priority updates across other classes would be awesome. Unlike melee, just 2 or 3 battles with a mega are usually sufficient to get a strong idea of where things stand. --Miked0801 09:10, 20 June 2009 (UTC)

You can also set RUNONLY=NANO, but alas then no priority battles are fought, just random ones. And I set the battlecount to 1000, so nanos will get their battles earlier. --GrubbmGait 10:51, 20 June 2009 (UTC)

  • setting RUNONLY=NANO is not a good idea. The results only count for nano, so is not counted at micro/mini/mega. Setting the battlecount lower does pay off however, except when the bigger rumbles are flooded with new versions. --GrubbmGait 11:00, 20 June 2009 (UTC)

It would be awesome if roborumble could execute smart battles with runonly parameters. For example, run prioritary battles but only among nanos. I don't think it would be such a difficult developement. --HUNRobar 14:33, 20 June 2009 (UTC)

I think the RUNONLY setting needs to be fixed -- as GrubbmGait pointed out, the battles don't get uploaded to the bigger size classes, which they should. I might put together a patch for this, but first I have some wave surfing to iron out... --Darkcanuck 17:41, 20 June 2009 (UTC)

About battlecount, one habit I have, is constantly making my battlecount set to just a little above the lowest value in the rumble. That seems to have nice results sometimes. --Rednaxela 16:03, 20 June 2009 (UTC)

All right, I think I've figured out how to implement the change I suggested, plus fix the RUNONLY problem. If enough people are ok with this, I'll put together a patched 1.6.1.4 client with the changes and send the patches to Fnl for future releases too. --Darkcanuck 18:21, 21 June 2009 (UTC)

Quick question: Would this only affect the 1v1 pairing order? I definitely support it there, but making your suggested change in Melee pairings would have a major impact on some bots' scores... And thanks as always for your RoboRumble efforts! --Voidious 19:33, 21 June 2009 (UTC)
Only at first. The game already does all class battles for roughly 50% of the numbers in nano now. That's why DustBunny has been slowly rising in APS - he does worse in nano only. --Miked0801 20:42, 21 June 2009 (UTC)
Well, right now a NanoBot gets 2,000 battles against the whole field, then a bunch against MiniBots and below, then MicroBots and below, then NanoBots. With this change, it would get some amount of battles against the whole field, then all NanoBot-only battles. Depending on the bot, I think it could have an impact. Then again, that's a quirky aspect of MeleeRumble that's tough to really regulate, anyway, so maybe it's not worth worrying about. --Voidious 21:51, 21 June 2009 (UTC)
Ah, good point, I hadn't considered melee. It does use a different routine for preparing battles though, so it's possible to patch 1v1 without affecting melee. Looking at the melee routine, it does create an interesting mix of battles -- I didn't realize that there were all-nano matchups. That must skew the scores depending on the ratio of battles a nano gets with mixed sizes versus all-nano? --Darkcanuck 02:15, 22 June 2009 (UTC)
I don't think it is a bug, i think it was made on purpose the way it is. Alas I don't remember the reason . . . --GrubbmGait 21:03, 21 June 2009 (UTC)

Mobile Devices

Who here things I'm crazy to try running roborumble on my Ipod Touch? :) --Rednaxela 02:01, 7 August 2009 (UTC)

I think you're crazy, but only because the iPhone OS doesn't have Java. =) I actually just looked into this the other day, lol. Was there some tech you were looking at to make it work? --Voidious 02:03, 7 August 2009 (UTC)

Actually, when you don't limit yourself to the app store, there is a port of JamVM w/ GNU Classpath. There's no JIT so it won't be fast... but it'll run.. or... almost:

Iteration number 0
Downloading rating files ...
Downloading participants list ...
Downloading missing bots ...
Removing old participants from server ...
Preparing battles list ... Using smart battles is true
Prioritary battles file not found ...  
java.awt.AWTError: Cannot load AWT toolkit: gnu.java.awt.peer.gtk.GtkToolkit
   at java.awt.Toolkit.getDefaultToolkit(Toolkit.java:607)
   at robocode.security.RobocodeSecurityManager.<init>(Unknown Source)
   at robocode.manager.RobocodeManager.initSecurity(Unknown Source)
   at robocode.control.RobocodeEngine.init(Unknown Source)
   at robocode.control.RobocodeEngine.<init>(Unknown Source)
   at roborumble.battlesengine.BattlesRunner.initialize(Unknown Source)
   at roborumble.battlesengine.BattlesRunner.<init>(Unknown Source)
   at roborumble.RoboRumbleAtHome.main(Unknown Source)
Caused by: java.lang.UnsatisfiedLinkError: Native library `gtkpeer' not found (as file `libgtkpeer') in gnu.classpath.boot.library.path and java.library.path
   at java.lang.Runtime.loadLibrary(Runtime.java:763)
   at java.lang.System.loadLibrary(System.java:662)
   at gnu.java.awt.peer.gtk.GtkToolkit.<clinit>(GtkToolkit.java:173)
   at java.lang.VMClass.forName(Native Method)
   at java.lang.Class.forName(Class.java:233)
   at java.awt.Toolkit.getDefaultToolkit(Toolkit.java:583)
   ...7 more

It seems that both:

  1. RobocodeSecurityManager is foolish and tries to query the default graphics toolkit even if run from the command line
  2. The install of GNU Classpath is foolish and thinks it has GTK avaliable when it's running in command line only

Were it not for one or the other of those issues, it'd probably run just fine... Now can i fix this I wonder... --Rednaxela 02:13, 7 August 2009 (UTC)

Wow, that would be really cool. But I'm a wuss and I'm afraid to jailbreak and void my iPhone's AppleCare; I was thinking more from the perspective of writing an iPhone app myself that could do it. Good luck, though, I'm curious to hear your results! --Voidious 02:17, 7 August 2009 (UTC)

Aha! I got around that stupid issue! I just had to pass the VM -Djava.awt.headless=true... When you see rumble uploads with a user of "RednaxelaIPod", that'll be it :) --Rednaxela 02:32, 7 August 2009 (UTC)

Ouchie, building the robot database this first time sure isn't fast... --Rednaxela 02:39, 7 August 2009 (UTC)

Sadly, while it mostly ran, it didn't quite work:

java.lang.ArrayIndexOutOfBoundsException: 1
   at java.util.concurrent.CopyOnWriteArrayList.get(CopyOnWriteArrayList.java:27
5)
   at java.util.AbstractList$1.next(AbstractList.java:354)
   at robocode.battle.Battle.updateBullets(Unknown Source)
   at robocode.battle.Battle.runTurn(Unknown Source)
   at robocode.battle.Battle.runRound(Unknown Source)
   at robocode.battle.Battle.run(Unknown Source)
   at java.lang.Thread.run(Thread.java:743)

...

java.lang.NullPointerException
   at robocode.battle.Battle.updateBullets(Unknown Source)
   at robocode.battle.Battle.runTurn(Unknown Source)
   at robocode.battle.Battle.runRound(Unknown Source)
   at robocode.battle.Battle.run(Unknown Source)
   at java.lang.Thread.run(Thread.java:743)

...

Exception in thread "sul.Pinkbot 1.1" java.lang.NullPointerException
   at robocode.peer.RobotPeer.setEnergy(Unknown Source)
   at robocode.peer.RobotPeer.setEnergy(Unknown Source)
   at robocode.peer.RobotPeer.run(Unknown Source)
   at java.lang.Thread.run(Thread.java:743)
Exception in thread "serenity.moonlightBat 1.17" java.lang.NullPointerException
   at robocode.peer.RobotPeer.setEnergy(Unknown Source)
   at robocode.peer.RobotPeer.setEnergy(Unknown Source)
   at robocode.peer.RobotPeer.run(Unknown Source)
   at java.lang.Thread.run(Thread.java:743)

Though if a workaround can be found for those issues (most likely bugs in the version of GNU Classpath in question. May not apply to newer versions, or a patched robocode might be able to work around it), it may work. Not sure that'll be possible though. Oh well... --Rednaxela 03:54, 7 August 2009 (UTC)

Rumble Computers

I'm curious, what kind of computers are people using for roborumble? Voidious and Simonton seem to really have some nice ones I think. Personally the computers I have avaliable to run rumble consist of:

  • Intel Core 2 Duo T5300 (1.73GHz), 2GB ram, laptop, usually avaliable
  • AMD Athlon XP 2000+, 512MB ram, often avaliable
  • AMD Athlon XP 2400+, 1GB ram, rarely avaliable

I would also have my wimpy iPod were it not for one little bug preventing successful running of bots in the VM installed there. Sometimes... I'm really tempted to get a brand new shiney quad core monster for the sole purpose of storing backups, and running RoboRumble/Research :) --Rednaxela 03:41, 16 August 2009 (UTC)

I have:

  • MacBook 2.0 GHz Core Duo, 2 GB RAM, and I can run two clients on it safely. It's actually 3+ years old, but it holds up pretty well.
  • AMD 2000+, 768MB RAM (Ubuntu 8.10), which I've been running a rumble client on almost 24/7 for a while now.
  • Just yesterday, I ordered a MacBook Pro 2.8 GHz Core 2 Duo with 4 GB RAM. So that will more than double my total rumble power, I think. =)

I wonder what Corbos had, I couldn't have kept up with him if I tried when he was rumbling recently. I've also been tempted many times to spring for a quad core Linux box just for Robocode, but so far I've been able to resist... (mmm, quad core...) --Voidious 03:55, 16 August 2009 (UTC)

Haha... I unfortuantely can't run two clients on my Core 2 Duo laptop... the CPU just plain gets too hot. What I really want to see though, is MXThreadBean timing in a new robocode version, which would make it safe to enable the option to allow robots to run concurrently, when in rumble... Then one could take advantage of a 8 beast when running melee rumble with only one client... --Rednaxela 04:29, 16 August 2009 (UTC)

My client runs on an old laptop -- all it does is run roborumble + meleerumble:

  • Intel Pentium M w/512MB ram, 100% available

Occasionally I press my main machine (MacBook Pro, Core 2 Duo, 3GB) into service, but mostly I need it for work or running RoboResearch (that's what it's doing right now, in fact). It's also where the development rumble server lives, so this client is normally pointed at that anyway. Running both I was able to keep ahead of GrubbmGait for awhile until Voidious et al made our upload counts look puny by comparison. --Darkcanuck 04:26, 16 August 2009 (UTC)

My client is a 2.66 Mhz Pentium4 (single-core) with 1GB. There was a time that I ran roborumble on a 400Mhz laptop with 128Mb using robocode 1.0.7. Btw, Darkcanuck, are you able to remove (or rename) those somewhat embarrasing contributions of GrubbmGait_15x and GrubbmGait_17x at the very low bottom. --GrubbmGait 11:08, 16 August 2009 (UTC)

Wow, 2.66 MHz!!?! :-D J/k. --Voidious 15:04, 16 August 2009 (UTC)
It's done. I merged GrubbmGait_15x into GrubbmGait_154, and removed the lone battle from GrubbmGait_17x (the upload still shows up, merged into GrubbmGait). --Darkcanuck 17:38, 16 August 2009 (UTC)

When I used to run it, I ran it on my Intel Centrino 2 P8400 2.26GHz with 4GB of RAM (but with 32bits OS it only sees 3GB) laptop, which I used for other works too. The time I run I only use FF to surf the web, but now I need to use Photoshop, Eclipse, LAMP and one virtual machine almost all the time so it doesn't available anymore. » Nat | Talk » 11:31, 16 August 2009 (UTC)

I run my client in a Pentiun Dual Core 1.86 GHz, and I usualy run 2 clients, but right now I'm using it to test several new versions of my robot, so the rumble has to wait...--Navajo 14:36, 16 August 2009 (UTC)

Lower Rank Limit?

Wouldn't it be reasonable to remove bots that gets negative ELO Rating (i.e elvbot.ElverionBot(ELO: -191.0) & ry.Worst 1.0 (ELO: -365.3) ) ?

They almost always score 0% and uses cpu-time while they contribute with near to nothing. --Rsim 15:37, 16 August 2009 (UTC)

Whoops

Err... sorry my client reactivated some old melee bots due to old data lying around, seems they crave pairings now... plus a large number of recently released melee bots... means sloooow melee right now... --Rednaxela 05:10, 10 September 2009 (UTC)

Iteration

I wonder if this happen to everyone or just me. When I run RoboRumble client, both 1.6.1.4 (which upload to darkcanuck's server) and 1.7.1.4 (which I use to test locally because it is faster), I found that after iteration 80 things got slower. I mean with 1.6.1.4 I can run 100 iteration in 4 hours, then another 4 hours it can run only 50-60 iterations. SO I tend to restart my client every 80 iterations. (and cool my hard disk down a little, it is around 80 degree Celcius when I run it for 150 iterations) Do anyone, especially who run RoboRumble overnight, have similar situation? --Nat Pavasant 03:10, 11 October 2009 (UTC)

I've found that 1.6.1.4 still seems to have some long-term memory leakage actually. Not as bad as older version, but enough that it's still worth not using the ITERATE option actually. --Rednaxela 03:54, 11 October 2009 (UTC)

JVM Crash

After running the RoboRumble client (1.6.1.4) for about 12 hours, my JRE crashed. I have not been able to reproduce the problem, but I thought I'd post about here in the event it has happened to someone before. I've submitted a bug report to Sun and saved the logs [Here]. It may not be relevant, but this happened when evaluating Krabb.sliNk.Garm 0.9u. --Christopher.Hilla 22:06, 7 February 2010 (UTC)

Huh... that's a very peculiar bug. I doubt there's anything more to it other than there being a bug in the JVM though, assuming this isn't something you're able to repeat reliably. --Rednaxela 22:48, 7 February 2010 (UTC)

Contributor's Sum

I started contributing less than 30 days ago, yet my "All time" and Monthly sums don't match. My "All time" sum is much larger than what I've actually contributed. Even though my point tally is sort of like keeping score on "Whose Line Is It Anyway", it's worth mentioning :) --Christopher.Hilla 06:49, 9 February 2010 (UTC)

The monthly and 30-day totals are based on battles, whereas the all-time list is number of uploads. For a 1v1 battle between megas there is only one upload, but for smaller classes there will be one upload per size class (e.g. two nanos will generate 4 uploads). Melee is even more complicated since a full set of pairings for 10 megas will yield 45 uploads, even more if there are smaller bots in the mix. So your all-time uploads should be greater than the sum of the battles you've contributed so far. --Darkcanuck 07:41, 9 February 2010 (UTC)
Thank you for the explanation. --Christopher.Hilla 19:49, 9 February 2010 (UTC)

Incomplete robot list

Hey Darkcanuck, I 've been running a few rumbles..and noticed I am missing a large number of bots " can't file specified file". Any Suggestions? Thx.. oh , I should mention Last night while Running an iteration envolving DR3.06 , Portia and 8 other bots. Something probably on my end screwwing up results for 1 battle. Not sure how, or why, but just one battle. -Jlm0924 20:06, 27 April 2010 (UTC)

DogmanSPE

I just ran a few tests to try and figure out the discrepancy between 1.8.0 and 1.8.2 of DrussGT, and now I can't repeat 1.8.0's results against DogmanSPE. I then decided to try a few rounds against Diamond 1.5.5, and it seems he also isn't living up to the scores he's getting on the rumble against DogmanSPE. The only thing I can think of is that it may be different versions of Robocode, as I'm doing my testing with 1.7.2b2. Any ideas? Maybe Dogman gets smart only at certain times of day? --Skilgannon 15:06, 9 May 2010 (UTC)

Eh, DogmanSPE is a major troublemaker and always has been. It sometimes doesn't crash, but when it does crash it crashes for every subsequent round of the battle. This makes it's results always unpredictable. It doesn't affect overall APS too badly because of the sheer number of bots, but it's probably the single worst offender in the rumble. Issues with DogmanSPE have been pointed out before. Personally, I'm in favor of removing DogmanSPE. --Rednaxela 15:25, 9 May 2010 (UTC)

So am I. Anyone object? --Voidious 17:28, 9 May 2010 (UTC)
Done =) --Skilgannon 19:17, 9 May 2010 (UTC)

LRP

Perhaps we should have a section on what the LRP does. I mean most of use from the old wiki know but... Anyway in the words of Paul,

May seems a stupid question, but, what is the blue line, and the 4 quadrants?
I can guess that on the X axis there is the expected score and on the Y the difference between expected and actual score.
But what the blue line is supposed to mean? -- [[Simonech]]

You are correct on the X axes and Y axes scales - the Blue line is a best-fit straight line through the data. if it has a negative
gradient slopes downhill (left to right) I believe it would indicate that the movement for the bot is stronger than the gun. A
positive gradient would indicate a good gun is being let down by relativley poor movement. -- [[Paul]]

Chase-san 04:55, 4 July 2010 (UTC)

Rumble use of 1.7.2.1 Beta?

Guys, do you think that Robocode 1.7.2.1 will be good enough to replace the very old 1.6.1.4 for RoboRumble? If not, what must be done? --Fnl 22:06, 29 June 2010 (UTC)

The main problem I have with 1.7.2.1 is that objects from the data section in jars from inside zips at least do not get extracted to the cache. I also have problems with duplicate robots in the repository (mostly only ones on the development path). — Chase-san 00:57, 30 June 2010 (UTC)
Huh, I was under the impression that the data files were extracted properly still. (note, it's technically not a robotcache anymore). Also, what do you mean by duplicate robots, and is this different than older versions? --Rednaxela 05:30, 30 June 2010 (UTC)
Well, I haven't noticed any new problems myself, but I've yet to have a chance to test very extensively, but I plan to on what is a long weekend coming up here. I'm going to clear my localhost rumble server and re-run testing with the latest version. Does anyone thing it would be good if I opened up the 1.7.2.1 testing rumble server to the world so others could speed up the process of testing it? --Rednaxela 05:30, 30 June 2010 (UTC)
I think it would be good. I am more than willing help you. If not due the stability of internet connection of my server I would open a testing server ages ago (well, my server located at my school in the university's infrastructure which is like twenty years old or more). Just a side note, 1.7.1.x runs at least five times faster than 1.6.1.4. --Nat Pavasant 12:51, 30 June 2010 (UTC)
Okay. I will wait a little with releasing 1.7.2.1 so you guys get the chance of testing it a bit more. 1.7.2.1 is "just" replacing the Jikes compiler with ECJ + some bug fixes compared to 1.7.2.0. I have started working on the 1.7.2.2 with some new features + I try to improve the speed with loading/updating the development paths, which are loaded extremely slow compared to the .jar files. :-) --Fnl 19:57, 2 July 2010 (UTC)

Alright, I just moved the bots that are known to not work in 1.7 for "WontFix" reasons out of the participants list. Anybody mind? I kept them on the bottom of the pages with links to the SF ticket that's relevant. --Rednaxela 20:46, 4 July 2010 (UTC)

Okay, I now have a rumble test server facing the world on [[1]]. It should be secured and fast enough now, I think. Feel free to run 1.7.2.1 Beta clients on it for testing! :) --Rednaxela 21:56, 4 July 2010 (UTC)

I have my roborumble and meleerumble client pointed to your server now. --Nat Pavasant 13:29, 6 July 2010 (UTC)

Red, do you have any comparing method planned? The melee is almost completed, though it should need more before becoming stable. --Nat Pavasant 14:48, 12 July 2010 (UTC)

My current comparing plan is:

  1. First look for obvious outliers (things missing from high ranks, things at extreme low ranks)
  2. Then import the 1.6.1.4 rumble and 1.7.2.1 rumble results tables into a spreadsheet, sort each by name, line them up beside eachother, create a column for rank difference and PL difference, then sort by those to find outliers.
  3. Test any outliers. Determine if it's a bug to report, or the bot doing something that it shouldn't

Currently, 1v1 rumble has two outliers at the bottom. I already reported the bug for one and it was fixed, and I think the other is breaking due to using File IO without using the proper robocode streams. About to do the spreadsheet with melee :) --Rednaxela 02:13, 13 July 2010 (UTC)

The biggest melee outliers by APS. Positive numbers mean improvement:

Rank change APS change
jab.DiamondStealer 5 145 20.07
gg.Squaraus 0.6 -72 -14.08
com.syncleus.robocode.Dreadnaught 0.1 49 9.40
stelo.Spread 0.2 62 6.34
bayen.nut.Squirrel 1.615 48 4.36
stelo.Moojuk 1.3 10 3.95
stelo.SoJNanoMelee 1.1 -23 -2.95
exauge.GateKeeper 1.1.121g 21 2.37
sgp.JollyNinja 3.53 18 2.12
lrem.micro.SpikeSoldier 0.4 19 2.00
sample.TrackFire 1.0 -0 -1.87

I suspect DiamondStealer may have had bad results on the main rumble server, and I suspect the APS differences below 3% are probably a fluke. Maybe some of the larger are too, but I plan to test Squaraus and Dreadnaught anyway. --Rednaxela 05:16, 13 July 2010 (UTC)

Dreadnaught also had bad results in the beginning (not complient to 1.5.4 I believe) --GrubbmGait 08:49, 17 July 2010 (UTC)

.NET

Is it possible to submit robots written in .NET to RoboRumble? --dserbia 15:18, 15 July 2010 (UTC)

No, not at the moment. We have to make sure Robocode 1.7.x reach the really stable phase before. And I think we should wait for jni4net to reach at least beta status, for Mono support. Many of high-contributing RoboRumble clients are running under either Linux or OSX. --Nat Pavasant 15:38, 15 July 2010 (UTC)

Thanks for the quick reply! --dserbia 17:21, 15 July 2010 (UTC)

Personally, I'm not sure I'd ever be in favor of allowing them into the rumble. There would then be bots in the rumble that Mac/Linux users could never test against (and there are a lot of us), and it would also make us dependent on .NET users to contribute battles for those bots when we post new versions of our own. But I do sympathize - if/when more .NET Robocoders come onto the scene, things could change.

Keep in mind that all of this stuff is community driven, though it may all seem very "official" to a newcomer. Anyone is free to setup a competition / tournament / rumble that includes .NET bots if they want! We've had tournament style competitions like the MiniBot Challenge and Twin Duel in the past, Robocode has APIs that make automating battles for tourneys fairly easy, and Darkcanuck posts the source to his rumble server somewhere (Nat or Rednaxela would know). And there's always Challenges you can compete in, too.

--Voidious 17:38, 15 July 2010 (UTC)

I think Pavel Savara has plan for suporting jni4net, which is the core of .NET Robocode, on Mono, which can be run under Linux and Mac OS X. I don't think there is a valid reason for not supporting those when that day come. --Nat Pavasant 06:57, 16 July 2010 (UTC)
Yeah, it could work out, but there's still lots of room for issues. How stable is Mono? Will Mono add behavior discrepancies to .NET bots? Or will Java bots hit subtle behavior discepancies when battling .NET bots on Windows? (Either of those could be very hard to debug for the bot author.) Will .NET bots run much slower when running through Mono? Will that count against them in terms of CPU time/skipped turns? We've seen plenty of issues just from different JVMs / CPUs / OSes, so adding Mono and .NET to the rumble definitely requires caution. Or it may work perfectly smoothly and we'll all coexist peacefully. =) --Voidious 15:18, 16 July 2010 (UTC)
As long as one isn't doing GUI or calls to native libraries (not allowed for bots anyway), I'm pretty sure Mono is quite accurately compatible, however my concerns are around the fairness issues. x86 Java is reasonably uniform in terms of what parts of the code get optimized how well by the JVM. As I understand, Mono's performance is pretty reasonable overall, however since it's coded from scratch it's optimization characteristics are notably different I think, which makes cpu fairness somewhat tricky for having MS .Net and Mono running the same rumble. If one looks at this, it's fairly old, but shows large performance differences going either way in terms of different types of operations. That exact issue only becomes larger between the JVM and .NET, so I really do think it makes sense for .NET to be kept to a separate rumble (possibly on the same server but separated like the different formats?). --Rednaxela 19:23, 16 July 2010 (UTC)

New Rumble

Rednaxela and I have been talking a bit, and we have put considerable thought into a new 1-vs-1 rumble. Unlike our current 1-vs-1 rumble, this would have no data saving, be single round battles on a 800x800 field. Also considering having the game 'white out' names in a battle to just numbers, so there can be no hard-coded specific bot data saving either. I personally feel this kind of 1-vs-1 competition would allow for much greater competition from non-adaptive movements, and lower the bar considerably for newer robocoders. But not too much, since obviously, Crazy would never be able to win over DrussGT.

We also considered non-random starting positions (for greater fairness) and a 4 bot single round melee on the same field size.

Chase-san 16:16, 24 July 2010 (UTC)

Yeah, I think this would be a really neat thing to have. Now, I don't think it should replace the old-style rumble, but I think it would make a very interesting competition, and perhaps much less daunting for newcomers. And to clarify on what the "white out" of names, it means robot names that bots would see would just be like "1", "2", "3", or perhaps "robot 1", "robot 2", "robot 3". Now, that does require a modification to Robocode, but I think it would probably be relatively simple and quick to get implemented. (I say it should be simple because Robocode already alters names when multiple of the same bot are in a battle, so I'd imagine the same code path can be reused, so just a matter of changing it for a certain rumble flag) --Rednaxela 16:27, 24 July 2010 (UTC)

We have put togeather a seed of 60 robots, please keep in mind this is just a seed, and you can add your own robots when it gets started. This is just to round out the robots so that we can get some meaningful rankings.

abc.Tron 2.02
ags.micro.Carpet 1.1
ags.polished.PolishedRuby 1
apv.Aspid 1.7
apv.NanoLauLectrik 1.0
apv.ScruchiPu 1.0
ary.nano.AceSurf 1.2
axeBots.Musashi 2.18
axeBots.Okami 1.04
chase.pm.Pytko 1.0
cjm.Che 1.2
cx.Lacrimas 1.36
cx.mini.Cigaret 1.31
darkcanuck.Gaff 1.50
davidalves.net.DuelistMini 1.1
davidalves.net.DuelistNano 1.0
dummy.micro.Sparrow 2.5
gh.GrubbmGrb 1.2.4
gh.GrypRepetyf 0.13
jam.mini.Raiko 0.43
jk.micro.Toorkild 0.2.4b
kawigi.robot.Girl 1.2
kawigi.sbf.FloodHT 0.9.2
kawigi.sbf.FloodSonnet 0.9
kc.micro.Needle 0.101
kc.micro.Thorn 1.252
kc.mini.Vyper 0.311
kc.nano.Splinter 1.2
kid.Gladiator .7.2
Krabb.krabby2.Krabby2 1.9o
mld.Moebius 2.9.3
mue.Hyperion 0.8
mz.NanoDeath 2.56
pe.SandboxDT 3.02
pe.SandboxLump 1.52
pedersen.Ugluk 1.0
pez.mako.Mako 1.5
pez.mini.Tityus 0.9.1
pez.mini.VertiLeach 0.4.0
ry.VirtualGunExperiment 1.2.0
rz.Aleph 0.34
rz.GlowBlowAPM 1.0
sample.Walls
sample.Crazy
simonton.beta.LifelongObsession 0.5.1
simonton.micro.GFMicro 1.0
simonton.micro.WeeklongObsession 3.4.1
simonton.nano.WeekendObsession_S 1.7
stefw.Tigger 0.0.23
synapse.Geomancy 15
techdude.Carruthers 1.2.6
tzu.TheArtOfWar 1.2
voidious.micro.Jen 1.11
vuen.Fractal 0.55
wcsv.Engineer.Engineer 0.5.4
whind.Strength 0.6.4
wiki.BasicGFSurfer 1.02
wiki.mini.Sedan 1.0
wiki.nano.RaikoNano 1.1
zyx.micro.Ant 1.1

Chase-san 17:36, 24 July 2010 (UTC)

1.7.x rumble eventually...

Hey guys... I kind of think I dropped the ball on working on getting 1.7.x to main rumble before... but I still think it would be good to eventually. Now... unlike before (where the rumble server was a little VM on my machine that I sometimes had to shut down for lack of memory), I have a more proper server I could use for rumble. I probably won't get around to setting this up till after winter holidays (have plans with family), but just wanted to ask, how much interest in this is there? Best Regards. --Rednaxela 07:29, 21 December 2010 (UTC)

I am still interested in. Right now (well, not really, since I am not actively writing robots), my movement predictor have flags for 1.6.1.4 or 1.7.1.3+ movement. Since rumble runs 1.6.1.4 but 1.7.x+ runs much faster. Also, the whole rumble would take less time to stabilised (but it would put more load to server, since client can run, as Pavel claimed, ten times faster. --Nat Pavasant 11:36, 21 December 2010 (UTC)

I'm still interested. Kept meaning to ask you why the test rumble server went down but never did. Do you still have all the results, or would we start from scratch? Anyway, it would be good to be current. A lot of new people get 0 scores because they use new APIs, plus there's the performance benefits. --Voidious 14:00, 21 December 2010 (UTC)

Yeah, I do still have the old data, but I'm not sure we want to use it. Presumably we'd want to test now with the very latest version, and mixing the results in this of multiple versions in this test could cloud results. --Rednaxela 14:59, 22 December 2010 (UTC)

Alright! I have a new rumble test server up and running at http://rednaxela-robocode.dyndns.org/rumble/. For ease of use, rumble configuration templates are linked from it's main page. Also, if you want, here's a convenient zip of all the bots currently in rumble: [2]. This one is located on a more proper type of server, rather than a virtual machine on my laptop. So... time to get testing 1.7.2.2 and prove it stable? :) --Rednaxela 06:37, 19 January 2011 (UTC)

Hm, I thought I had this page "watched" but it looks like I've missed a lot... Is the new server running the same code and database schema as the current one? --Darkcanuck 06:43, 19 January 2011 (UTC)
Yep, same code and database schema. The only changes I made were to the front html page, and changing the accepted version of robocode to 1.7.2.2. --Rednaxela 06:57, 19 January 2011 (UTC)
I found my old laptop, I will set it to run, I might tweak the CPU constant up a little too, since it is a older/slower machine. (I will just set it in my closet, away from hustle and bustle). — Chase-san 09:32, 19 January 2011 (UTC)
Nice! I'll start running my clients against this again. Thanks Rednaxela! Btw, the superpack .zip is gone from RoboRumble/Starting With RoboRumble. I'll make a new one with the above archive unless you could just repost the old one. --Voidious 14:47, 19 January 2011 (UTC)
It turns out both my laptops puked at some point while I was out (one locked up while loading rankings, the other I forgot to make a loop and iterate was NOT). Not sure if you still have to do iterate not, and make a script loop, but I did. Only my main system had kept running the rumble. But I restarted both of those now. — Chase-san 02:36, 20 January 2011 (UTC)
As far as I know, all the memory leaks have been fixed. I had my client run over the whole day and the memory use plateaued nicely with no issues whatsoever. --Rednaxela 02:39, 20 January 2011 (UTC)
Ah well, I will switch them to use ITERATE then. Thanks. — Chase-san 02:59, 20 January 2011 (UTC)

Is there anything that we're waiting on for the switch to the 1.7 rumble rankings? I'd love to help make this upgrade happen. Ncj 16:45, 4 April 2011 (UTC)

I know Rednaxela was investigating some of the outliers from the test server's rankings. I'm not sure there was really anything "stop ship" though. I think the rankings look pretty good. Rednaxela, what do you think? --Voidious 20:38, 4 April 2011 (UTC)

Well, here's the "Table of Anomalous Results" I made when the 1.7.2.2 rumble first stabilized:

Old Rank Old % New Rank New % Rank Diff % Diff ABS % Diff ABS Rank Diff Explanation
CharlieN.Omega.Omega 1.03 319 55.29 774 9.19 -455 -46.1 46.1 455 tracker item 3207405
NDH.GuessFactor 1.0 787 0.42 578 39.54 209 39.12 39.12 209 uses post-1.6.1.4 functions
jeremyreeder.Vincent 2011.12.09 788 0 645 33.68 143 33.68 33.68 143 uses post-1.6.1.4 functions
Krabb.krabby2.Krabby2 1.9o 71 73.63 528 43.12 -457 -30.51 30.51 457
gg.Squaraus 0.6 714 26.3 786 4.62 -72 -21.68 21.68 72 tracker item 3207405
elvbot.ElverionBot 0.3 785 0.72 751 19.75 34 19.03 19.03 34 uses post-1.6.1.4 functions
TJ.Exupery 1.39 403 50.02 174 63.3 229 13.28 13.28 229
zh.UnderDog 0.0.2 782 4.31 762 14.11 20 9.8 9.8 20 1 on 5 exception in TacticalAdvisor in 1.6.1.4
reaper.Reaper 1.1 357 53.1 194 62.6 163 9.5 9.5 163 crashes on some clients in 1.6.14 but not 1.7?
sm.Devil 7.3 517 44.13 351 53.45 166 9.32 9.32 166 crashes on some clients in 1.6.14 but not 1.7?
xiongan.Xiongan 1.1 709 27.05 614 36.22 95 9.17 9.17 95
ag.Gir 0.99 762 13.25 739 22.26 23 9.01 9.01 23

The case of "CharlieN.Omega.Omega 1.03" was tracker item 3207405, and has been fixed in 1.7.3.0 Beta. Unfortunately I've yet to investigate the other results.

Some of positive results I suspect may have been from using some Robocode 1.7 features perhaps (i.e. some new events). The case of Krabby2 I in particular mean to investigate because that result looks quite odd. Some other positive changes like Devil seem strange to me, because Devil predates Robocode 1.7 by a huge margin, yet seems to do better than in 1.6 by a notable margin. I've yet to be able to figure that out either. Anyone have any theories or can look into these cases?

Anyway, presuming these issues can be worked out, what do people think of moving straight to 1.7.3.0 perhaps? The difference relative to 1.7.2.2 is small so I don't see it as that high risk, and if it turns out there were issues, then we're no worse off than if we had spent the same effort to test 1.7.3.0. Thoughts?

--Rednaxela 04:45, 5 April 2011 (UTC)

SmallDevil crashes on my clients (dunno about others), so his 1.6 score is probably just wrong. I'm probably OK with moving to 1.7.3.0, but I'd prefer to just go with 1.7.2.2 since that's the one we've actually spent so much time testing. The difference being small could support either decision. =) 1.7.3.0 is still beta, too. --Voidious 12:44, 5 April 2011 (UTC)

Well, of course some testing of 1.7.3.0 needs to be done, but that should be possible to complete before it's out of beta. I think it's changes from 1.7.2.2 are definitely seem small enough to me that it's not like we need to do a full rumble run again. --Rednaxela 13:23, 5 April 2011 (UTC)

The three bots 'NDH.Guessfactor', 'jeremyreeder.Vincent' and elvbot.ElverionBot' all use post-1.6.1.4 functions, no further investigation needed for them. SmallDevil indeed occasionally crashes, why it does not do that on 1.7.2.2 I don't know. I believe the same goes for 'reaper.Reaper'. Friday I will have some time to check other differences. Is there a possibility to check all '2100-0' scores on the server for a specific bot? A bad client on the moment of submission has a permanent influence now. On the old server (years ago) these bad results faded away with each new battle (0.7*0.7*0.7 etc). --GrubbmGait 18:35, 6 April 2011 (UTC)

I suspect SmallDevil just crashes due to some corrupt saved data, tho I haven't investigated it. I've seen the same with other bots, like TheBrainPi. I wiped my robotcache today so I'll watch out for SmallDevil battles from me with any new bot releases. --Voidious 00:30, 7 April 2011 (UTC)

Interesting. Tomorrow I think I'll be able to look into these more, especially Krabby2 I think. --Rednaxela 05:00, 7 April 2011 (UTC)

For reaper.Reaper and sm.Devil just check their scores against ad.last.Bottom. There is the possibility that these bots still could crash in 1.7, as it only happens occasionally. Is there a way to reject/ignore 21xx-0 scores (if other scores are available) and scores way outside the normal range? That would be like an aspirine for the ranking. See example below. --GrubbmGait 14:18, 7 April 2011 (UTC)

BATTLE DETAILS FOR "reaper.Reaper 1.1" VS "abud.ThirdRobo 1.0" IN GAME "roborumble"

|% Score |% Survival |score |bullet dmg. |survival |score |bullet dmg. |survival |Battle Time             |Submitted by 
|                    | reaper.Reaper 1.1           | abud.ThirdRobo 1.0          |
|90.622  |100.000    |5991  |3233        |35       |620   | 614        | 0       |2010-05-31 12:57:18:835 |cberendt@178.63.10.12 ver.1.6.1.4 
|90.927  |100.000    |5963  |3215        |35       |595   | 590        | 0       |2010-02-09 07:49:56:713 |Miked0801@207.170.253.83 ver.1.6.1.4 
|90.044  |100.000    |5960  |3209        |35       |659   | 657        | 0       |2010-01-06 02:27:33:942 |darkcanuck@24.85.46.67 ver.1.6.1.4 
|89.834  |100.000    |5956  |3200        |35       |674   | 667        | 0       |2009-01-15 16:51:12:182 |darkcanuck_154@24.85.45.250 ver.1 
|90.435  |100.000    |5928  |3187        |35       |627   | 622        | 0       |2008-09-10 15:49:11:187 |ABC@import ver.1 
|88.993  | 97.143    |5862  |3179        |34       |725   | 643        | 1       |2008-09-21 12:06:41:943 |ABC@import ver.1 
|0.000   |  0.000    |   0  |0           | 0       |2103  |   0        |35       |2010-11-09 13:21:30:209 |Voidious@76.15.102.168 ver.1.6.1.4 
|0.000   |  0.000    |   0  |0           | 0       |2101  |   0        |35       |2010-07-04 13:50:23:859 |Voidious@76.15.102.168 ver.1.6.1.4 
|0.000   |  0.000    |   0  |0           | 0       |2101  |   0        |35       |2009-10-13 23:34:34:50  |Voidious@173.50.167.108 ver.1.6.1.4 
|0.016   |  0.000    |   1  |0           | 0       |6179  |3395        |35       |2008-10-29 04:30:08:727 |darkcanuck@24.85.45.250 ver.1

I think we could wait for Version 1.7.3.1 is out of beta and use that for its bug fixes (or just go with 1.7.3.0). It is kind of silly to still be using a version for the rumble that is nearly 3 years old, it is unlikely we are ever going to get a `perfect` version. Also might be able to eventually get .Net bots into the rumble. — Chase-san 11:09, 5 July 2011 (UTC)

I agree, we should do a switch. The new Robocode runs a lot faster, meaning faster updates on battles for any given number of clients. Moreover, there are a lot of bots that don't work on older versions due to painting methods etc that have been added in the latest Robocodes. Also, using the latest Robocode version makes it easier for beginners to contribute to rumble as well as submit their bots. I haven't seen any widespread change in scores, the problems all seem to be due to individual bots having bad programming practices or being buggy. About .Net bots in the rumble, AFAIK there is no way to run .Net reliably on Linux, which is how I (and many others) run rumble clients, but having an up-to-date client is certainly the first step. What think ye, people? --Skilgannon 12:28, 5 July 2011 (UTC)

So, lets do it already. If we just don't do it it will never happen! After that we can maybe try to talk again about trimming the participant list, mostly because I think it really needs to happen eventually. Cause as things are going in 2046, we have 2900 participants, we still haven't trimmed the list and are running a robocode version from 2016. ;) — Chase-san 15:18, 5 July 2011 (UTC)

My only concern about 1.7.x on RoboRumble is server load. 1.7.x runs much faster than the older versions. About .NET, I think Pavel is still working on jni4net on Mono, but it is not stable nor usable as far as I know. And the main .NET is still in alpha. It would be long time before .NET robot can run on RoboRumble. --Nat Pavasant 12:46, 5 July 2011 (UTC)

Agreed, let's do it. The test rumble server results looked very good. But we might have to wait until Darkcanuck says he has time to oversee it in case there are any issues. Agree on leaving out .Net and disagree on trimming participants list... 4x the participants in 35 years when we have 1,000,000x the processing power doesn't seem like a problem. =) And maybe by then we will have moved to Robocode 2 anyway. :-P --Voidious 16:09, 5 July 2011 (UTC)

You forget that robots get more complex as time passes. ;) — Chase-san 17:28, 5 July 2011 (UTC)

Let's switch. What about the current results, are they kept or dismissed. I think with the switch from 1.5.4 to 1.6.1.4 we kept all scores. Only the problembots listed above need some attention in that case. Next to this switch, maybe it's time to create a new division: Robots, and move the 'Extend Robot' ones to that place. It depends a bit if can be distinguished if a bot is a Robot or an AdvancedRobot though. It should help in two ways, first Robots finally have their own ranking and can be compared, second the 1v1 becomes a bit less crowded (although that does not seem to be very annoying right now). --GrubbmGait 17:42, 5 July 2011 (UTC)

I am pretty sure the number of "Extends Robot" robots is proportionally small. Being generous, maybe 120 of them at most. — Chase-san 17:47, 5 July 2011 (UTC)

Like the previous switch, the server should temporary accept 2 versions of clients for the rumble. And I agree with Voidious that Darkcanuck should have some time to monitor the switch. And regarding a 'Robots' division, many robocoders start with ExtendsRobot as part of a (school)project or because they think it is easier :-P . Teamrumble and Twinduel do not have many participants either, although it is true those are very different than 1v1 --GrubbmGait 17:54, 5 July 2011 (UTC)

So it sounds like we're ready to add 1.7.3.0 to the approved list? The change is easy for me and I should have enough time to keep an eye on the server for the next week or two. 1.6.1.4 will still be allowed for a short period, but I think it would be good to disable it soon (because of bot compatibility). And if 1.7.3.1 looks good it shouldn't be such a big deal to add it in soon (ie. don't need to wait 2 years!). --Darkcanuck 19:23, 5 July 2011 (UTC)

I say go for it. Having looked at the code, I know you use an array for the version checks. (should put that into the config!) — Chase-san 20:08, 5 July 2011 (UTC)

I'm definitely in support of moving to 1.7.3. Also, a couple comments on other aspects of the conversation:
- About .NET, I think that even when there are no technical issues to running it with the roborumble client, it would be best kept in it's own rumble league. I say this because I doubt there cpu constant fairness could really be achieved.
- About the concern of server load that Nat mentioned, I know that my 1.7 roborumble test server held up just fine with a peak of many more clients than the main rumble server has ever had to cope with (thanks to Chase-San :)), and there wasn't really a problem. I'm not sure what Darkcanuck is running, but my test server is running on a quaint 128MB RAM VPS even (fitting mysql on that, while having some headroom, requires some performance compromises)
--Rednaxela 00:27, 6 July 2011 (UTC)

I think the bigger issue in that case was bandwidth rather then server load, them having used up something around 40% or 60% of its pipe (I forget). In case your wondering, I ran around 16 high end i7s, each with 4 clients, all at once. — Chase-san 01:12, 6 July 2011 (UTC)
The bottleneck on my 1.7 test server wasn't bandwidth. Note, my VPS has successfully pulled from kernel.org at over 600Mbit before, and reliably gets over 100Mbit. So... bandwidth is not an issue. I believe the bottleneck was to do with the MySQL server. The main ways to fix it would involve it eating more memory really. --Rednaxela 05:19, 6 July 2011 (UTC)
Re: server load, I'll keep an eye on it. The bottleneck is the huge size of the database: I suppose if a lot of old pairings were pruned out that would probably help (indexes would stand a better chance of not getting swapped out to disk). Or if I had a bit more memory I'd put the entire pairings table into Redis and the rumble server would be screaming fast. Right now there are just too many bots! --Darkcanuck 02:10, 6 July 2011 (UTC)

I'm going to shut down the server in an hour or two, backup the database and then bring it back online with 1.7.3.0 uploads enabled. Then we'll need someone to test it... --Darkcanuck 02:10, 6 July 2011 (UTC)

It's done. Any volunteers to test this? --Darkcanuck 03:59, 6 July 2011 (UTC)
YAHOOOOO!!!!!!:) I run 1.7.3 client soon --Jdev 05:09, 6 July 2011 (UTC)
Done. --Jdev 05:17, 6 July 2011 (UTC)
And NDH.Guessfactor 1.0 gets its first real score ... --Darkcanuck 05:29, 6 July 2011 (UTC)
Starting a 1.7.3.0 client here --Rednaxela 05:19, 6 July 2011 (UTC)
Started a client here, just one though. Might setup a deticated box at some point. Also in a month or two we can probably flush all the older 1.6.1.4 pairings. — Chase-san 12:52, 6 July 2011 (UTC)
I'll start a client tonight. About flushing the 1.6.1.4 pairings, why should that be done. That just means that a lot of valid information and info of older botversions is lost. Removing older invalid results (210x-000x scores) is usefull, but that should be an automated task. A point of discussion could be whether versions that are retired for more than x years (f.e 3 years), should be moved to another (inactive) database. In that case the 'running' database will not grow that fast. --GrubbmGait 13:39, 6 July 2011 (UTC)
Don't worry, I'm not planning on flushing anything. But it might be helpful to re-release some of the bots which used to crash under 1.6.1.4 so that they get a valid ranking (and give predictable results to their opponents) if their authors are still around or someone is willing to update the packages. Purging old pairing data (for retired bots only) might be nice for performance reasons, but the downside is that you lose comparison information so I'm not keen on that either. People have donated *years* of cpu time to build up the current database! --Darkcanuck 13:51, 6 July 2011 (UTC)
Maybe a good "flushing old data" strategy, would when calculating pairings, don't count battles from old versions if there are at least 3 battles of the pairing with the new version? It seems to me this would: 1) not be too hard to implement, 2) would phase out results that were bad with the old version without needing re-release of bots, and 3) be rather seamless in that it wouldn't immediately be tossing old version results out. --Rednaxela 14:11, 6 July 2011 (UTC)
Good points all around, what was I thinking? — Chase-san 14:13, 6 July 2011 (UTC)
It's an interesting idea but I have two concerns: 1) only a handful of bots have different results between versions, it seems wasteful to throw out perfectly good 1.6.1.4 battles (considering some pairings have a high APS variance) when we could just re-release the problem bots; 2) the pairing update is done incrementally so it currently doesn't query battle data, which could be fairly expensive to add. --Darkcanuck 14:34, 6 July 2011 (UTC)

I'm running some battles locally and noticed scoring 80% vs YersiniaPestis. This is actually with 1.7.2.2. I checked recent rumble results vs other top bots and found a 24% score vs DrussGT, which seems pretty unlikely: [3]. I'll look into it sometime, but in case anyone else gets a chance first... --Voidious 02:24, 18 July 2011 (UTC)

No problems with YersiniaPestis for me, well, GresSuffurd still loses. Could it be an OS/Java version problem? --GrubbmGait 08:08, 18 July 2011 (UTC)
I've always seen wild scores with YersiniaPestis during testing (probably due to my gun's affinity for YP's movement) but I don't think that battle is too unusual. There's a similar score vs Pris: [4] (using 1.6.1.4). Checking its other close matchups, there's a lot of variability, often > 20% APS. --Darkcanuck 16:32, 18 July 2011 (UTC)

New version of robocode is out, apparently with things to speed up rumble running as well. Shouldn't we update the rumble to allow 1.7.3.2? — Chase-san 01:32, 24 August 2011 (UTC)

I don't see any major core changes listed since 1.7.3.0, should be safe to try. If you want to run a test client and are willing to double-check the results, I can update the server to allow 1.7.3.2. Also glad to see my patch get released so quickly! --Darkcanuck 04:13, 24 August 2011 (UTC)
I suppose I could. Battles I have run in roboresearch seem to be highly similar to their rumble counterparts. I'll run it under the name Chase1732, so that they can be easily removed if issues occur. — Chase-san 04:40, 24 August 2011 (UTC)
It's ready. No need to change your name: internally you get a unique id based on your username+ip+version combination. --Darkcanuck 04:55, 24 August 2011 (UTC)

Client on Google Apps Engine

I wonder, since the latest release of Google Apps Engine (1.5.0), Google add support for Backend, which allow process to run as long as you like. I think it is possible to port the RoboRumble@Home client to run on these environment; that would speed up the process. But I think we need to check if this is applicable with Google Apps Engine TOS and throttle the client enough so it doesn't use more than its free quota limit ;-) Just an idea! --Nat Pavasant 01:42, 11 May 2011 (UTC)

Well, I'm pretty sure that even if you can run a process as long as you like, they don't guarantee a constant amount of available CPU, which is important for robocode. Of course, if in the future, robocode switches to using the less granular APIs that track actual cpu time used by given threads, then this would probably work. It's not really the same, but as a note, you can already run robocode on Amazon's EC2, which does guarantee a level of CPU allocation. Chase-san briefly tried this once in the past.--Rednaxela 02:51, 11 May 2011 (UTC)

But, of course, unless you already host something on EC2, the EC2 isn't "free" ;-) --Nat Pavasant 12:19, 11 May 2011 (UTC)

I've been running two instances of the client on the cheapest server offered by linode ($20/mo). I'd run more, as it's quad core, but it only has 500MB of RAM and is thus memory bound. I keep the box around in case I need to pretend my traffic is coming from the US, but for a dedicated RoboRumble box would EC2 end up being cheaper? Ncj 16:52, 17 May 2011 (UTC)

As an aside, I'd only recommend running one client with 500 megs of RAM, as I think other people use and test with 512. I remember the default was 512 in robocode.sh but still 256 in roborumble.sh at one point - not sure if that's been fixed. Pretty cool you do that, though! --Voidious 21:44, 17 May 2011 (UTC)
Aside from agreeing with Voidious about wanting 500mb per client, I'd question whether the CPU allocation on linode is stable. Those are virtual machines shared with other instances, and while there are "four cores" that it has access to, it's not as if it's exclusive access. I'd be very concerned that your clients are causing bots to skip turns whenever other instances on the machine get busy. --Rednaxela 00:29, 18 May 2011 (UTC)
If Linode CPU's allocation isn't stable, then isn't the EC2. They are based on the same virtualization technology, Xen. I'd be more concerned if it is KVM or OpenVZ virtualization system. --Nat Pavasant 14:29, 18 May 2011 (UTC)
It would all be a matter of how they configure it (i.e. cpu caps). EC2 is marketed for high performance computing purposes. I'm pretty sure caps cpu usage of instances so people never get more than they pay for, and also so it's possible for all instances to hit the cap without fighting for CPU. In contrast, I'm pretty sure Linode is markets their instances more for hosting purposes, and so I'd expect them to manage their caps differently such that when there are free cpu cycles an instance may use more than it would be able to if all instances were busy. Basically, the underlying technology (Xen) is the same, but as I understand, the type of service is fairly different. That said, I've personally not used either service, but from what I've read, this is my understanding of the two. --Rednaxela 23:22, 19 May 2011 (UTC)
Is there a way for me to check this? Somehow compare the % of skipped turns on my client to someone else's? Ncj 15:36, 18 May 2011 (UTC)
Unfortunately no (at least not without modification to robocode). Actually... I kind of wish that Robocode would report "percent of turns skipped" to the roborumble server.... --Rednaxela 23:22, 19 May 2011 (UTC)

Frequency of Submissions

I've been submitting a lot of updates to my robot lately. Is there any etiquette on how often is too often? I just want to make sure I'm not annoying anyone. I do expect my frequency of updates to slow dramatically soon. Thanks. Skotty 19:17, 13 June 2011 (UTC)

My view's pretty simple: as long as you're not delaying other bots from reaching a stable rating, release as often as you like! A bot without full pairings takes precedence over a bot under 2k battles, but a bot's rating isn't really stable until around 2k battles. So if my bot's at 800 and you keep releasing new versions, I'll never get to 2k. Personally, if this happened to me, I'd just temporarily exclude you on my clients, which seems fair to both of us. Pulling my weight in RoboRumble clients tends to alleviate any potential guilt, too. And FWIW, I've certainly done daily updates for long stretches (eg, Diamond/Version History and Komarious/Version History). --Voidious 19:42, 13 June 2011 (UTC)

Robocode Repository Package

Hello everyone. Robocode Repository is still down for longer time again. Is there a possibility adding a packge with all of its actual robots from the rumble? The roborumble-superpack-20110123 package is great stuff, but does not contain all of the actual Robocode Repository robots. Thanks! --FromHell 14:48, 20 June 2011 (UTC)

I zipped up all the JARs in my RoboRumble dir pretty recently, should have most of what you need: roborumble_bots.zip (69 MB). It also has lots of versions of many bots, so you might want to just grab what you need... It would indeed be good to update the superpack sometime. --Voidious 14:56, 20 June 2011 (UTC)

Great! Thank you very much. Yes, updating the superpack would be good, because sometimes the stabilization of the rankings needs some time, because not every contributor has all bots, especially the ones missing from repository. Maybe I will update my local rumbleback the next days, after finishing my actual development on CHCl3. Normally I would publish the new superpack, but the only problem is, that my webspace is from a friend of mine and is very limited and he wants to change the provider in the next few months. So it's no good idea publishing it on "my" server. But it would work for a while. --FromHell 18:50, 20 June 2011 (UTC)

I addad a package to my webspace with all missing Robocode Repository bots (for every category). roborumble_missing_bots_20110621.zip (275 KB). It would be great if someone would update the superpack with these bots and making it available for the whole community. Or should I add my link to the Starting With RoboRumble page at this wiki? --FromHell 09:54, 21 June 2011 (UTC)

ImWithStupidTeam

There seems to be a problem with ImWithStupidTeam. Executing the rumble will be canceled with the following message: "Aborting battle, could not find robot: pedersen.ImWithStupid 1.3". I had a closer look to the bot and there was still a jar file in the jar file. Taking the inner jar, adding the file ImWithStupidTeam.team to the jar, renaming the new jar to pedersen.ImWithStupidTeam_1.3.jar should just do the trick. Should this team be removed from the team rumble participants list? Found an entry to 1v1 list that there was also removed a bot, because it was wrong packaged. What is the right way? Thanks for your answers. --FromHell 10:30, 21 June 2011 (UTC)

Thanks to MN for taking action at this point. --FromHell 21:00, 3 July 2011 (UTC)

Running the rumble

I apologize but it is unlikely i'll be able to run the rumble this week. Forecast is all very high heat and high humidity. I personally do not want to run my system in the jungle. In fact so much so I cannot even use it, as it is off. — Chase-san 19:50, 18 July 2011 (UTC)

I was following a policy of running clients non-stop until the rankings stabilize whenever I mess with the participants list myself. --MN 02:08, 22 July 2011 (UTC)

I will be running the rumble almost all day Monday, Tuesday, Wednesday and Thursday at the current rate, also at night (almost every day). So feel free to release more often or use rumble for more developmental builds. Please put my extra cpu time to good use. — Chase-san 17:49, 8 September 2011 (UTC)

Increase BATTLESPERBOT roborumble parameter

I think, that it is time to increase BATTLESPERBOT, beacuse current value gives ~2.4 battles against every but and this amount is too small to stabilise ratings
I increase BATTLESPERBOT to 4400 battles (~ 5.3 battles against each bot) and suggest to all do the same thing. Or counter arguments:) --Jdev 05:54, 11 August 2011 (UTC)

I don't think it's necessary. The latest updates to the server cause it to send out priority battles for any low matchups 50% of the time (which overrides the client's battles-per-bot processing). That should ensure pairings fill up fairly evenly. Extra priority is also given to pairings with < 2 battles, those are requested 100% of the time. I would be inclined to increase the 50% instead, if we all agree that even pairings are most important (I think they are). --Darkcanuck 06:05, 11 August 2011 (UTC)
I'm Sorry, but because my english i can not understand following statements:
* "send out priority battles for any low matchups 50% of the time"?
* "Extra priority is also given to pairings with < 2 battles, those are requested 100% of the time."? Anyway, 2 battels is still too low. May be increase this value?
* "that even pairings are most important"

--Jdev 07:14, 11 August 2011 (UTC)

I briefly described the new system here: Darkcanuck/RRServer/Updates. Every time you upload a result, the server decides whether or not to send your client a priority battle request (a single pairing that needs another battle). These priority battles get run first, then your client worries about the bots with battles < BATTLESPERBOT. Right now the server always sends a priority battle request if it sees a pairing with < 2 battles. Otherwise, 50% of the time it will send a request for a pairing that has a low battle count. The other 50% of the time, your client will choose what to run next. We could either increase the percentage or increase the < 2.
"that even pairings are most important" = trying to get all pairings to an equal number of battles is most important (in my opinion). --Darkcanuck 07:41, 11 August 2011 (UTC)
Thank you, Darkcanuck! Now all things is clear for me. --Jdev 08:06, 11 August 2011 (UTC)
There was a discussion on Talk:Darkcanuck/RRServer/Query. The only disadvantage I see with making the priority system too deterministic is with the ELO and Glicko-2 rating systems. They are sensitive to battle ordering and work better with random pairings. Right now they are both stabilizing slower than the APS rating system. --MN 13:52, 11 August 2011 (UTC)
This kind of thing is why in the past I've argued in favor of having a rating system which is essentially APS, except uses Glicko-2 to estimate the score of missing pairings. I say this because it seems to me the only advantage of Glicko-2 over APS is in the incomplete pairing scenario, and one should be able to get the best of both worlds by just using Glicko-2 for estimation of missing pairings. --Rednaxela 14:18, 11 August 2011 (UTC)