Starcore AI Pack 4.0 Reboot

[doot]

Stock's been slowly dying down, and it's kinda become a HS, Popup, SNS fest.

Whenever I see something like this posted, 90% of the time it's from someone that only builds IRL.

Also, fixing up or making slight tweaks to half the bots in an existing AI pack is easier than making something like 100 or 200 new robots from scratch.

To be fair, though, you do have the obvious bot types that have huge advantages in certain metagames, but not with others. Stock has the HS/SnS/Popup flood with a few other bot types, but you do have other bot types in metagames that have their obvious upsides, like with IF's Popups/VS'es/Flippers, DSL 2.2's efficiency curve and Hammertraps/VS'es/rammers and Retooled's similar weapon efficiency curve, only shallower. Then again, I haven't really done much of other metagames to really have a true view, so I may not be the most reliable.

I do agree with you in that re-tweaking robots to be better than before is easier than remaking everything from scratch, though.

[Starcore]

IMO, Starcore was never a max efficiency AI pack in my eyes. While it definitely has effective designs in it, and it certainly helped move stock forward, it's greatest strength has always been the diversity of bots. There are lots of bots in SC you won't see in another AI pack.  Yes bot types aren't all equal, but it's like that in every meta and has been like that since the start of ra2, and as long as rebuilds are built to the best of the design while still resembling their predecessors, bitching about the meta shouldn't be an issue.

Geice is very much on target here. Yes teams like my personal Spin Doctors were built to be as efficient as I could back then to the predominate theme of HS (except of course AW Logic Probe as a form of FS due to limitations of what you can do at Antweight levels) across the teams and bots being effective was important AND so was exposing various building techniques, styles and yes, types of glitches that allowed, even more, bot types and examples.

Before Starcore v3 Antweight, Beetleweight and as I recall Superheavyweights didn't exist in Stock AI packs. We standardized the specs and gave lots of examples.

If Popups, HS and SnS are currently dominate then yes it is worth taking effort to develop and present Anti-Popups,  Anti-HS, and Anti-SnS.

Review overhead hammers and bring up to current teams like Team Jabberwocky. I remember working on a Scorpion style bot to be Anti-Popups (with wedges and/or ramplate front and caster armor).

Work on Rammer AI to make bots like Warhammer more effective with better pull back and ram code.

SnS are vulnerable to flippers for OOTA and while difficult to AI in Stock, possibly Clamp Bots.

I asked ACAMS for out-of-the-box thinking and that is why you see a full team of side hammers and a team of anti-popups.

You have seen the SHW Locutus of Borg with its Anti-Gut Rippers. This can be done with more standard weight solutions. And there are AI packages that can handle multiple Smartzones for different effects.

When I asked Rejected for some out-of-the-box thinking we got bots like Neglected Waterbug (one of the first and likely the most famous juggler in RA2) and Clockwork Hydra with its piston thrusting VS solution that sought to solve the issue of opponent bots being too wide or having spinning weapons so the VS was not able to get close enough to the opponents chassis. And he called out to his friends and we got bots like Papa Roach and Lil Darling.

And you may feel otherwise but I have never felt we had top-notch AIed Pop-ups in the Starcore packs. We had some relatively nice pop-ups but you all know many have done better and that is something that is much easier to drive than properly AI.

I know some have taken the challenge of going against the Starcore packs in the internal Tournament mode and surviving an entire tournament as a test of overall toughness. I think we need more Anti-Popups and Popup Protection in the pack because about the only types of bots that have ever finished the Starcore Pack Tournaments have been very defensively driven armored Popups.

So, is there a reason and interest to teach latest techniques and show off current best-of-breeds in a Stock pack update?

Can 'you' build better? Are there things we can still teach?

And as Mr AS and Dreamcast mentioned, I am not looking to rebuild the whole pack. I am not looking to make it non-Stock. It was at an Alpha2 state and I would like to bring to you all a finished version of that pack. What bots are Alpha broken? (AI messed up? Steel armor that should be DSA? Etc) What bots are just too ineffective to provide reasonable build examples and need to be replaced? Being still in Alpha I was opening the door to like maybe 10 teams to bring the pack to a full 45 teams for tournament rotations thereby giving you all a chance to show off a team or two of your greatest and brightest.

Starcore

[Squirrel_Monkey]

I feel like downloading the latest version and just giving my feedback on every bot. Is there a download with all the bots in or do I have to extract them from the AI teams?

[Starcore]

The packs are available at: https://gametechmods.com/forums/index.php?action=downloads;sa=myfiles;u=5/

I recommend installing Starcore V3 then replacing the /AI teams with the /AI team in the Starcore v4 Alpha2 pack.

Note: It is an Alpha and not a full pack release.

Starcore

[Squirrel_Monkey]

The packs are available at: https://gametechmods.com/forums/index.php?action=downloads;sa=myfiles;u=5/

I recommend installing Starcore V3 then replacing the /AI teams with the /AI team in the Starcore v4 Alpha2 pack.

Note: It is an Alpha and not a full pack release.

Starcore

I was more hoping someone had extracted the bot files from the teams so I could open them in the lab. But that's just me being lazy.

[Starcore]

If you are wanting them prepped for Botlab review with all their names changed back from Bot0, Bot1, Bot2, etc to their actual names, I am pretty likely to be the one most able to do that since I have Botlab and Archive versions of all the bots.

But currently no, I don't have anything like that prepped for 'lazy' bot builders. 

Now, if anyone else has already done that legwork you can hope they notice this forum thread and pipe up.

I respect that you and others are willing to help, providing it doesn't take you too much effort to dig in, so I as the requester will see what I can do this weekend...

Starcore

[090901]

Just from looking at the bindings, all the weapon ids need to be fixed using apanx's bot exporter (exe or installer).

[Starcore]

Yeah that was a short cut back then to not worry about binding numbers matching the weapons and even on bots I had right, they fell out of sync as updates and rebuilds occurred.

I didn't worry about it much as long as I had enough IDs to be at least equal to the number of weapons because that is the point that the bot AI switched from Attacking to Shoving. If you didn't care much about when that happened as long as it didn't happen too early it didn't matter much that the numbers matched the Weapon binding numbers.

Shove was a common choice for pushing an opponent into a pit or out of the arena or by locking their weapons and causing a Failed to Move TKO. Better than winning by, "Let's run away some more!"

It did matter when one did want to switch to an alternate response like fleeing when its actual weapons were removed.

If people are concerned that the information is right, we can do a project when the rebuilds and tunings wind down to match binding numbers to actual weapon part numbers. Hope Apanx's bot exporter is a decent shortcut tool for that these days. It was a PITA back then.

Happy to have your help with that when it comes time,

Starcore

[Starcore]

Hmmm.... My work's McAfee Anti-Virus scanner reports that Apanx's Bot Exporter in the GTM Downloads area has a Trojan and won't let me download it.

Any discussion about this being a false alarm?

Would be good to recompile the .EXE so it doesn't report as such. Often even a small change will induce enough of a difference to remove the false alarm (if it is a false alarm).

Starcore

[Badnik96]

i personally would love to help finish this pack off and maybe submit a team or two.

also i'm pretty sure the trojan is a false alarm, given i downloaded it ages ago (and have definitely scanned my computer since) and nothing has cropped up for me.

[Sage]

Always down to help out!

[Starcore]

Yeah, I didn't think it was a real Trojan. More likely a false alarm. That is why I mentioned how to recompile it to get rid of the issue.

I will check it from home since my work won't let me download it to protect me.

Starcore

[Silverfish]

I'd be happy to do what I can. I'd love to take a whack at billy goat gruffiest or alien queen.

[090901]

Always down to help out!

have you finally perfected your rebuild of jabberwock? 

[Meganerdbomb]

Honestly? I don't know.

Stock's been slowly dying down, and it's kinda become a HS, Popup, SNS fest.

I feel maybe creating an DSL Standard pack would be best, maybe using Mr. AS's rebalance pack, because Stock is kind of stagnant currently.

Actually,  stiltbots currently dominate the meta. Starcore AI needs some stiltbots. Also, I've never been the best stock builder, but as a fan of V3 back when it was released, I'd love to contribute.

[Starcore]

When you say Stiltbots, I assume you mean bots like Mome Rath of Team Jabberwocky in Starcore v3 based on Bumbler by Corporate Goon.

Glad to see that style of bot has come into its own. As I recall the issue with that kind of bot was designing the legs wide enough to not be taken out by a HS, SnS or pretty much any "wide" bot.

Starcore

[Meganerdbomb]

When you say Stiltbots, I assume you mean bots like Mome Rath of Team Jabberwocky in Starcore v3 based on Bumbler by Corporate Goon.

Glad to see that style of bot has come into its own. As I recall the issue with that kind of bot was designing the legs wide enough to not be taken out by a HS, SnS or pretty much any "wide" bot.

Starcore

Yeah, the 140cm square extender got accepted as a legit stock part a few years back (it was a hidden stock part that is 140 cm and weighs 5kg), and its main value has been making stiltbots actually practical. They're still weak vs wide has/sns, but HS becoming more compact actually helps them.

[Dreamcast]

When you say Stiltbots, I assume you mean bots like Mome Rath of Team Jabberwocky in Starcore v3 based on Bumbler by Corporate Goon.

Glad to see that style of bot has come into its own. As I recall the issue with that kind of bot was designing the legs wide enough to not be taken out by a HS, SnS or pretty much any "wide" bot.

Starcore

Yeah, the 140cm square extender got accepted as a legit stock part a few years back (it was a hidden stock part that is 140 cm and weighs 5kg), and its main value has been making stiltbots actually practical.

Just don't try to build an anti-VS stiltbot. The journey is painful and the pay is in Zimbabwe dollars.

[09090901]

When you say Stiltbots, I assume you mean bots like Mome Rath of Team Jabberwocky in Starcore v3 based on Bumbler by Corporate Goon.

Glad to see that style of bot has come into its own. As I recall the issue with that kind of bot was designing the legs wide enough to not be taken out by a HS, SnS or pretty much any "wide" bot.

Starcore

Yeah, that's the style of bot. One way to cheat the max sphere to make the bot wider is using bursts to mount your legs so you can move them inwards to mount your drive and supports, and then move them outwards past the max sphere when you're done.

You wouldn't happen to have any of the old rebuilds sent way back, would you?

[Mr. AS]

You wouldn't happen to have any of the old rebuilds sent way back, would you?

Not him, but Virus Bomb was nice enough to send me some old rebuilds of his.