My Feedback

[Vezzra]

I'm basically thinking that, once Influence is in the game at roughly the same time we should half all the bonuses (so that it's 0.1, or 1 per 10 population)

Yep, that's what I've been thinking too - get the influence stuff in and then redo the boni. Although I have a much more aggressive toning down in mind - roughly to 25% of what they are now (meaning, 0.05, or 1 per 20 pop). The flat (pop independent) boni need to be adjusted accordingly of course.

However, with the influence stuff in place just halfing the current boni (and probably a more aggressive toning down of those ridiculous boni e.g. the BHG grants) might be sufficient. We'll see I guess

[IamZeke]

Two remarks:

First, the problem isn't so much that ships are too cheap, but that resource output increases too fast/too much. The current boni granted by techs, buildings, specials are simply far too high. You start with a PP output of ~10, which ramps up to several K in the later stages of the game. That is insane and impossible to balance. This issue has already been pointed out repeatedly in the past, we just didn't get around to redo all those boni, as that's one hell of a task. But it has to be done at some point.

The rise in PP may be partly to blame, but the AI still tries to overcome opposition strength by reaching parity any way it can. With the emphasis in the game on low slot number hulls it has to rely on stacking up ships. It would use bigger ships if it could to achieve opposition parity if they were available. Bigger ships mean few ships and less decisions per turn.

Reducing PP is fine as part of the problem to solve. So would creating larger slot hulls.

Another would be to get the AI interested in other research options. While I'm spending PP trying to add planet sensors, lighthouses, terraforming, gaia shifts, etc, it is just building hulls. It never tries to optimize its planets. When I'm taking their territory I stay bogged down trying to fix the systems I take. It will drop a colony ship of just about any race on any planet. The AI makes dog planets and simply goes for sheer numbers to make up for it. It's just drilling down to hard numbers to achieve superiority. It sees numbers and piles on to meet them with quantity. There are no planetary optimization algorithms.

Second, regarding how the AI places its fleets, contrary to the impression you apparently got, the AI currently has no idea of topology and can't identify chokepoints (AFAIK). It just assigns some kind of "score" to each system to decide how much defensive fleet power to station there. Which leads to what you have observed, ship stacks at a lot of systems where they aren't really needed, or don't make much sense. It also leads to the problem of fleets of troop ships sitting at systems where you can easily pick them off, instead of more protected areas.

The perks of an unfinished game...

It definitely sees opposition numbers at roadblocks because it keeps piling on to match what it sees in the next blocked system.

As for not seeing topography elsewhere I would agree. It just piles up decoys and scout ships everywhere. All those decoys and scouts each require a decision made each turn, even if they can't move. Piling up 50 decoy ships that do almost nothing for defense only adds to the AI decision burden each turn. And needing a scout for every system in order to see wider is yet another decision burden that a planet sensor would solve.

[Kassiopeija]

What do you want with big hulls early on? You don't have the necessary production to complete such a ship anyway - esp. if you're also suggesting to generally increase hullcosts as well... Let's say it takes 20 turns to construct a single such ship for Egassem - and 60 turns for Etty - then the Egassem cruiser has 40 turns to wipe the floor with a virtually completely defenseless Etty empire, and if they manage to take the planet where the ship is being constructed for Etty it will never finish and all previous invested production goes down the drain... Least to say they will never release a single fighter & can be conquered by a single hostile attack cruiser.
In a compact generated galaxy such an option would be a "I-win-button" to anyone who exploits the most out of it.

The ability to construct a midgame/endgame-hull (like Gravitating/Titan) would completely obsolete any existing smaller shiphull that their is, and their techs along with it.

The cost of the things available in game always need to be balanced against the output that can be achieved at that stage of a game. If it is too high then a game becomes boring because turns default to empty turns. If it is too low you get too much stuff which means turns become too long esp. in microing always the same repetetive tasks. This is the case in my games around turn200 but then the game is already won & simply a mop-up & instead of building new ships I terraform & turn GGs into planets etc

The early game however is very well balanced. There was, and is, an eternal debate & recognition about the fact that there's too much prod in the game for years, (and how esp. how strong Nasc & Automated early on are) but all that's done so far was to add more specials [like the cheesecomb one^^]

And I don't agree that the early shiphulls are useless. I've won a 300 system map with Etty with only using the basic Robotic Hull for warships. The only problem were planets on defensefocus, all other fights were won lossless due to Robotic Interface:Shields - which is extremely hard to make good use of early on but if you can survive pays off in the long run.

What we need [I will keep on stressing this point until you finally surrender!!! ] is a difficulty-setting administering handicaps to the player &/ bonuses to the AI [or vice versa for Beginner mode] and then, it will become increasingly IMPOSSIBLE to win combats lossless [= stuff gets taken out of game/ eliminated from memory] and each player can then go and find his own diff setting in order to keep the game challenging [ie. turning the mopup into a fight of survival]

[Vezzra]

It definitely sees opposition numbers at roadblocks because it keeps piling on to match what it sees in the next blocked system.

I know, but not because it has an actual sense of topography, but because it simply is programmed to counter big stacks by building its own big stacks in an adjacent system (AFAIK).

[IamZeke]

What do you want with big hulls early on? You don't have the necessary production to complete such a ship anyway - esp. if you're also suggesting to generally increase hullcosts as well... Let's say it takes 20 turns to construct a single such ship for Egassem - and 60 turns for Etty - then the Egassem cruiser has 40 turns to wipe the floor with a virtually completely defenseless Etty empire, and if they manage to take the planet where the ship is being constructed for Etty it will never finish and all previous invested production goes down the drain... Least to say they will never release a single fighter & can be conquered by a single hostile attack cruiser.
In a compact generated galaxy such an option would be a "I-win-button" to anyone who exploits the most out of it.

The ability to construct a midgame/endgame-hull (like Gravitating/Titan) would completely obsolete any existing smaller shiphull that their is, and their techs along with it.

The cost of the things available in game always need to be balanced against the output that can be achieved at that stage of a game. If it is too high then a game becomes boring because turns default to empty turns. If it is too low you get too much stuff which means turns become too long esp. in microing always the same repetetive tasks. This is the case in my games around turn200 but then the game is already won & simply a mop-up & instead of building new ships I terraform & turn GGs into planets etc

The early game however is very well balanced. There was, and is, an eternal debate & recognition about the fact that there's too much prod in the game for years, (and how esp. how strong Nasc & Automated early on are) but all that's done so far was to add more specials [like the cheesecomb one^^]

And I don't agree that the early shiphulls are useless. I've won a 300 system map with Etty with only using the basic Robotic Hull for warships. The only problem were planets on defensefocus, all other fights were won lossless due to Robotic Interface:Shields - which is extremely hard to make good use of early on but if you can survive pays off in the long run.

What we need [I will keep on stressing this point until you finally surrender!!! ] is a difficulty-setting administering handicaps to the player &/ bonuses to the AI [or vice versa for Beginner mode] and then, it will become increasingly IMPOSSIBLE to win combats lossless [= stuff gets taken out of game/ eliminated from memory] and each player can then go and find his own diff setting in order to keep the game challenging [ie. turning the mopup into a fight of survival]

Game balance is important, but when your idea of balance causes the game to simply stop working due to decision lag for so many ships then all you get a balanced game you can't finish.

There needs to be less decisions per turn for the computer to stop the lag. It's easy to see at the endgame that a full size galaxy with most planets populated but no real volume of AI ships that lag isn't being caused by planets. So what is left to cause the lag? Ships. The AI makes too many ships. Every AI ship requires a decision every turn. It's as simple as that. The AI is always going to drive to achieve numerical firepower in order to win. That leaves exactly two ways to reduce ships. Make them cost more to make and allow fewer hulls to achieve the necessary firepower. Combat is always an unknown variable. You can't use that as a guaranteed way to ensure fewer hulls. That would fail on several game settings, like Cautious and Turtle.

You might not like bigger and more expensive hulls. You may feel it blows game balance. But the game must first be able to operate to endgame without crushing lag that makes player stop the game before completion. The game you want to play takes a backseat to the game you can actually finish.

I stopped coding decades ago but I still understand decision trees. Too many AI decisions slow the game. It only gets worse the more complicated the decision tree, but the larger the decision tree the smarter the AI will respond. Unless you are willing to dumb down the AI response to reduce decision tree size then you have to reduce the overall number of decisions made. Fewer ships equals fewer decisions made without reducing the AI's ability to make complex decisions.

This is a game with minimal graphics to eat system resources and it still causes crushing lag on very good gaming computers. Not everyone has ability to buy a rig like mine. And it still makes me sit waiting a minute or longer to finish a turn at just turn 300. I've seen enough comments here to know that while everyone has a pet peeve or desire, the lag issue is mentioned so often it now is tossed out as a given. I'm as guilty as the rest wanting new features or game changes, but I also know that adding more detail to the game isn't going to make the lag issue any better. Lag needs to be dealt with to see what can be done. Otherwise just about any new interesting feature is going to slow the the lag down even more.

[MatGB]

The lag problem isn't, generally, AI decision making, it's effectsgroup processing.

Every object in the game that has an effectsgroup (most have several) has to be processed every turn, and because of the game structure being deliberately open so non-coders can easily adapt/change add content the effectsgroup processing can never be 100% efficient.

You're right that the number of ships is a problem, but not because of AI decision making, it's that all the hulls, and each individual part, has to be processed through at least twice because they frequently effect each other in some way.

[Kassiopeija]

I also raised the lag-issue multiple times but your solution wouldn't actually solve anything - if my current tolerance of lag allows the creation of a 500 system galaxy, and you generally reduce totalshipcount by 50% - then the only thing it'll ultimately mean is I can now create a galaxy holding 1000 systems, which will, again, bring me to the point of aborting a game.
Which, as of now, is somewhat pointless because without a proper diff setting the game is far too easy and cannot be lost anyway, resulting only in repetetive boring grind after turn200.
But someone else might start these games then we're virtually be back to square 1 - so maybe we should increase shipcosts by factor 100 because AFAIK the upper game limit of system generation lies at 5000!!!

Looking past I remember older versions of FO dozens of months ago were I couldn't even play out maps with 300 systems. So the game did definitely improve.

And it has nothing to do with my system because on this very rig I once played a Galactic Civilization 2 all-abundant game colonizing approx 1000 planets and build +95.000 single ships and had them all fly through a gigantic galaxy to gain military score under a MSBA. The problem with lag lies completely within the FO code itself.

[Kassiopeija]

You're right that the number of ships is a problem, but not because of AI decision making, it's that all the hulls, and each individual part, has to be processed through at least twice because they frequently effect each other in some way.

But there must be more at large than this. Because, even at the starting screen the cursor already becomes laggy & isn't updated so frequently. If I draw a circle with it I see the cursor jumping from places several centimeter apart, the common windows behaviour (which is retained in any other application - even emulators - I have here) updates, at least, 3-4 times more swiftly.
It really makes juggling the small arrows in options difficult - and as the gamelag (from ships) gets worse the cursor also worsens with it even more (until it's downright impossible to select a planetary focus correctly... ie. you basically have to backcheck everything you do....)

[IamZeke]

I also raised the lag-issue multiple times but your solution wouldn't actually solve anything - if my current tolerance of lag allows the creation of a 500 system galaxy, and you generally reduce totalshipcount by 50% - then the only thing it'll ultimately mean is I can now create a galaxy holding 1000 systems, which will, again, bring me to the point of aborting a game.
Which, as of now, is somewhat pointless because without a proper diff setting the game is far too easy and cannot be lost anyway, resulting only in repetetive boring grind after turn200.
But someone else might start these games then we're virtually be back to square 1 - so maybe we should increase shipcosts by factor 100 because AFAIK the upper game limit of system generation lies at 5000!!!

Looking past I remember older versions of FO dozens of months ago were I couldn't even play out maps with 300 systems. So the game did definitely improve.

Sometimes you just have to put hard caps in the game. If you raise the price of hulls then you help lag problems, but you can't then go stab yourself in the back by allowing an excess of other things that weigh it back down again. If that means you cut the number max systems then so be it.

In any case, I've already played larger maps and noticed at the end that lag drops once the AI doesn't have a lot of ships. Be it the actual ship counts, interactions between components, or whatever, the lag primarily comes from the AI having increasing numbers of ships on the map. More ships means more lag, regardless of the code reasons why.

As for monotony, almost most pro games cover that with planned and random events. The monsters in this game help that problem to a degree but not completely.

And it has nothing to do with my system because on this very rig I once played a Galactic Civilization 2 all-abundant game colonizing approx 1000 planets and build +95.000 single ships and had them all fly through a gigantic galaxy to gain military score under a MSBA. The problem with lag lies completely within the FO code itself.

Comparing pro games isn't a direct correlation. They have trimmed and closed code.



Ok, now please help me with a problem setting up the game. Somehow I have too many empty systems on my maps. Many times the number is half or more and too many empty systems seem to be at useful intersections. With the AI always playing soviet style numbers games and not wasting PP on making systems better it needs a bit of choke point strategy to compensate for that. Yet it seems I can never get the number of choke points down to a nice enjoyable level. I'm forever playing whack a mole with the aggressive hordes finding yet another way in.

I set the game to high planetary counts but it doesn't help. I go with spiral galaxies and low starlane settings to limit access. Nothing helps. What settings get the actual systems populated more. Not more planets per populated system, but more systems populated.

If there is no settings for that then there should be. Empty systems truly suck. Empty choke point systems suck more. Too high a number just causes whack a mole aggravation. I have a laundry list of issues with the game, but understand it's a work in progress and I also will have different objections than others do. So I've really tried to limit the things mentioned here to things that make me want to quit playing it. Lag is obvious and a concern for many. I don't know if the empty systems bother others as much as me, but if all I'm doing is rushing ships out to fill new holes in the line because the game likes to have a lot of holes then I can't enjoy a bit of empire building. Exploit is one of the 4X pillars and one of the most popular ones. I want to be able to grab a decent sized area, lock it down, and exploit it in order to strategically overcome the AI horde tactics.

[AndrewW]

If there is no settings for that then there should be. Empty systems truly suck. Empty choke point systems suck more. Too high a number just causes whack a mole aggravation. I have a laundry list of issues with the game, but understand it's a work in progress and I also will have different objections than others do. So I've really tried to limit the things mentioned here to things that make me want to quit playing it. Lag is obvious and a concern for many. I don't know if the empty systems bother others as much as me, but if all I'm doing is rushing ships out to fill new holes in the line because the game likes to have a lot of holes then I can't enjoy a bit of empire building. Exploit is one of the 4X pillars and one of the most popular ones. I want to be able to grab a decent sized area, lock it down, and exploit it in order to strategically overcome the AI horde tactics.

Try a cluster or ring setup instead of a spiral.

[MatGB]

As for the krill spawner tech, that also goes back to the dearth of internal slots. Internal slots are just too valuable real estate to waste on it. It might help to actually state what the stealth bonus is in the Pedia.

Yeah, I think we need to go over the Pedia and remove all the vague "you'll find out" explanations and replace them with specifics, but one of the reasons they're sometimes not specific is because they've been known to be changed and not updated, at one point the Fractal Hull had one description in the Tech, a second in the Hull entry and neither was correct, so the number of slots had been changed at least twice before it was updated, the Scattered Asteroid had a similar issue. We have, deliberately, gotten a lot better at keeping things up to date now, especially with the active translations going on.

Having said that, reading "internal slots are too valuable" is a really nice thing, until relatively recently internal slots were an irrelevence, when I last did a hull costings balancing pass I ignored the number of internal slots almost completely. That we've both increased the number of parts to make it a real choice and made some of the existing parts significantly more usable over the last couple release cycles is a really good thing—it's probable that we'll need to think about internal slot numbers and definitely recost ships taking them into account moving forward, especially with Fighters coming in next cycle.

Maybe have it spawn extra tasty krill that really draws krakens from a long distance. Then it would be great to use as a stealth weapon to keep the opposition busy with kraken attacks. Something like a 10 starlane range.

This is possibly a good idea, would certainly be more interesting, have Kraken actively hunt Krill. If someone wants to knock up a basic implementation that would be something I'd enjoy testing.

[Vezzra]

If there is no settings for that then there should be.

I agree. Currently there is no setting, at least in the game setup dialog, for the probability of "no star" systems (the galaxy age setting affects that to some degree though). You can however tinker with the probabilities for the different star types in the universe tables (found in universe_tables.py). That of course is definitely not an acceptable substitute for a proper game setup setting, which I plan to introduce (one of the many items on the ever growing todo list ).

Empty systems truly suck.

This apparently is very much a matter of personal taste. I tend to agree with you here, but others seem to prefer more empty systems. Originally we had far fewer of these deep space systems, their number has been substantially increased because there was a strong demand for that. Introducing a setting by which players can accomodate the probability of deep space systems to their personal preferences sounds like a good idea to me.

Another related problem that has already been brought up is that all systems which are not deep space systems should have at least one planet/asteroid belt. Currently you can get (especially on the low planet density setting) quite a lot of systems that have a star, but are otherwise empty. Making sure that each system with a star to also have at least one planet/asteroid belt will further cut down the amount of those empty systems.

[Geoff the Medio]

But there must be more at large than this. Because, even at the starting screen the cursor already becomes laggy & isn't updated so frequently. If I draw a circle with it I see the cursor jumping from places several centimeter apart, the common windows behaviour (which is retained in any other application - even emulators - I have here) updates, at least, 3-4 times more swiftly.
It really makes juggling the small arrows in options difficult - and as the gamelag (from ships) gets worse the cursor also worsens with it even more (until it's downright impossible to select a planetary focus correctly... ie. you basically have to backcheck everything you do....)

Make sure you have the game window selected when doing such a test, as it will drop the FPS substantially when it doesn't have the UI focus, in order to minimize background CPU usage. There are separate options for FPS limit for in and out of focus mode, in the options video section.

That said, if you're getting noticeably insufficient FPS on the main intro splash screen, then you've got some pretty bad video driver problems, and/or are using software rendering only. There's not really much rendering happening on that screen, other than a big image being blitted to the background each frame. It's quite possible for you to have a fancy new video card, but for it to have rather bad legacy support. FreeOrion uses rather outdated rendering calls for a bunch of rendering, and a 1 year old card is probably more interesting in making all the fancy new shader-based code work well than GL1.5 or 2.0 fixed-pipeline code...

If you have FPS drops in the game, eg. on the map screen, try going into the options screen and turning off starfield rendering and galaxy gas rendering, and perhaps also visibility circle rendering in the galaxy map screen. Also turn on the FPS indicator in the video tab to get some numbers. Preferably do this on a newish test build, as previous release have had FPS indicator calculation problems.

[Kassiopeija]

Make sure you have the game window selected when doing such a test, as it will drop the FPS substantially when it doesn't have the UI focus, in order to minimize background CPU usage. There are separate options for FPS limit for in and out of focus mode, in the options video section.That said, if you're getting noticeably insufficient FPS on the main intro splash screen, then you've got some pretty bad video driver problems, and/or are using software rendering only. There's not really much rendering happening on that screen, other than a big image being blitted to the background each frame. It's quite possible for you to have a fancy new video card, but for it to have rather bad legacy support. FreeOrion uses rather outdated rendering calls for a bunch of rendering, and a 1 year old card is probably more interesting in making all the fancy new shader-based code work well than GL1.5 or 2.0 fixed-pipeline code...If you have FPS drops in the game, eg. on the map screen, try going into the options screen and turning off starfield rendering and galaxy gas rendering, and perhaps also visibility circle rendering in the galaxy map screen. Also turn on the FPS indicator in the video tab to get some numbers. Preferably do this on a newish test build, as previous release have had FPS indicator calculation problems.

I have the game window selected. If I start a new map the FPS are shown as 59, but if I select the techscreen or production panel, it'll drop down to 29 (and stay there)

My vid drivers are up to date, it's an old card AMD HD 5700 running at a Phenom X4 965 system. I've got no clue about the rendering options, because I don't go into CCC and tinker with options I've really no clue what they do anyway, so I just install the latest updates and leave things at their default settings - hoping that a game may be able to adjust things on its own. Which seems to be the case in all cases except FO. I have a multitude of games installed and they all run fine. (With the exception of Distant Worlds Universe if creating a large map) I can run all games at ultimate/maximum graphical settings (eg Skyrim, HOMMV or AoW3) none of these games show any problems with the mousepointer whatsoever, and I wouldn't actually expect it from FO which doesn't even have excessive 3D graphics

I run FreeOrion_2016-05-19.4781a80_Test_Win32_Setup.exe. I followed your instructions and turned off all these renderings, but it won't help - I mean the drop occurs even in the techselection screen where the mainmap isn't shown at all...

Comparing pro games isn't a direct correlation. They have trimmed and closed code.

I also run likewise open-source games like Birth of the Empire or Battle for Wesnoth and my mousepointer is fluid there just like normal.

As for monotony, almost most pro games cover that with planned and random events. The monsters in this game help that problem to a degree but not completely.

Yeah that's what's needed. Monsters are more interesting in the early game, later none of them pose a threat anymore. Once I'm in touch with all AI and can observe how technologically developed they are, and if I can expand into their territory without much effort, I usually take a look at the graphs and if my productive & research output is bigger than all their output combined - I consider the game won and put it down.

With the AI always playing soviet style numbers games and not wasting PP on making systems better it needs a bit of choke point strategy to compensate for that. Yet it seems I can never get the number of choke points down to a nice enjoyable level. I'm forever playing whack a mole with the aggressive hordes finding yet another way in. I set the game to high planetary counts but it doesn't help. I go with spiral galaxies and low starlane settings to limit access. Nothing helps. What settings get the actual systems populated more. Not more planets per populated system, but more systems populated.

I don't understand what chokepoints have actually to do with planet-count ? Chokepoints are generated by the number of starlanes in conjunction with galaxy shape - try low + irregular to get most of these. You'll have an easier time defending your territory but for the AI it sometimes may cripple his game if his bottlenecked somewhere. A player can overcome this much more creatively.
I create such maps in hopes of being in a relatively bad to very bad starting position where I really can't do much and desperately find alternatives ways or researchpaths in order to get to a point of building myself up.

The AI benefits greatly from a homogenous map generation - like, put a planet on any system having a sun instead of creating 3 empty space system adjacent to a system which has 3 planets. Esp. in the early game because they aren't going to use supply-increasing mechanisms.

IIRC single-sun systems usually always have a multitude of planets at their disposal - it's usually double or triple-star systems which have catapulted mostly their planets out.

[IamZeke]

Try a cluster or ring setup instead of a spiral.

Thanks for the suggestion.

Yeah, I think we need to go over the Pedia and remove all the vague "you'll find out" explanations and replace them with specifics

Again, decades since I last coded in school, I remember my C- grade for a perfectly running program that fit all the demands asked, except one. Documentation. I barely had bothered because I assumed a perfect program spoke for itself.

He said if I had documented it I would have aced the class with a 4.0 grade. They were hyper about documentation. The reason given was that working perfectly or dying in an endless loop it could still be troubleshooted if documentation was solid. Solid documentation also made it easier to graft on other subprograms. One student got a higher grade than me and his program was buggy in a dozen or more places. But he documented like crazy and the professor was able to debug it fast. While I was irked by the hard grade strike I did at least come to understand the point they were making.

I learned one trick at least. Space and a marker to come back and work up documentation was always added while I was working on the code. I'd backfill documentation later once I got it working like I expected. Some kind of note should be added to all incomplete pedia entries. Even just some kind of marker like triple asterisks. If you put that in like that them you might even find some people willing to do that modest leg work for you, if they can't actually work on the game itself.

This is possibly a good idea, would certainly be more interesting, have Kraken actively hunt Krill. If someone wants to knock up a basic implementation that would be something I'd enjoy testing.

I do remember the comments about all monsters getting along, but in a single case it might be an interesting novelty. Right now a player has to go capture a kraken nest but if you could just lure wild ones you could have some real fun. Risky fun though because black kraken would like krill too and when the experimenters start releasing those then all players could be in trouble if you are spawning krill. Stealthed biowar, seeding mines outside your systems, and luring kraken could really be an interesting trio to ramp up a clandestine effort without too many big game additions. In some form or fashion the game already has all of the aspects needed. Clandestine war could be a great way to expand the game without overlaying an entire new system, such as trade or politics would require.

This apparently is very much a matter of personal taste.

The more you let the players adjust the initial layout the less startup objections they can complain about. If they unbalance the game the fault lies with them, not you. No need to try making one side happy while another complains because you feel you need to find a working balance. That you have already started down the "startup choice" road it only makes sense keep on with it.

I have already mentioned having an option for plentiful home systems and racial hull preferences, as examples to this kind of thought. These don't need you to decide if you think they are wise or unbalancing. Just add them as choice options. Then the player takes the blame if he doesn't like what happens.

I don't understand what chokepoints have actually to do with planet-count ?

It was just a guessing stab at trying to reduce the number of empty systems. It's clear now that is wasn't going to work. I have to try tinkering with the game or wait until it becomes a startup choice. Though I will try cluster and ring as Andrew suggested.