Data Mining
Well I am sort of late to the party but here is my two cents.
-First off I think Datamine seems to be on the up and up, they are looking for data from inside a real time strategy game, NOT stats in terms or palyer ranks, etc. Such data is hard to come by. Allegiance can offer it as the code is open thus they have the opportunity to modify the server and get people to play on it, allowing them to freely collect game play data. A server side only change would be best.
-Datamine, you need to realize that while we want to grow, we are a relatively small community and small player base. We have at most one game (of more than 20 people, 10 per side) going most of the time. It may not be worth your effort to mine data from this game.
-We already have our players splintered across multiple public servers. The best way to get players to play on your server is to offer a very low latency server with high reliability (i.e. it is up 99% of the time).
-For your server to be 'well met', you'd need to be listed on our public lobby, for that you need to talk to TigerEye and Pook. It would be good for you to run TAG as well so we continue to get the data we are collecting as well. Getting TAG installed means talking to TigerEye and Pook.
-Providing said server for only a month or two is likely going to be problematic unless we have a good idea of when it will and won't be up. Players need a good reason to play on your server versus other servers.
-I for one would like to see the code modifications you have made (well plan to make) to the server. Recognize that the code is not the easiest to navigate and that you will need a STABLE server. If your server becomes known for being buggy and crashing, people won't play on it.
Data collection wise I think the following (which is only a partial list) would make sense:
-Track time spent in each type of ship. It would be useful to know what percentage of flight time was spent piloting a bomber, or nan (scout equiped with a nanite repair gun), or int, etc.
-Track the amount of resources collected. For Allegiance that means He, so track how many miners are out, and the rate at which He is coming in. A graph of that data is going to be very very telling I think.
-Track how quickly the map is scouted. And then track how well it is rescouted, probed, and KEPT probed.
-Track what tech is bought when
-Track if people are following commander orders (can be a bit challenging to code up but you should be able to get at that at least a little)
-Track how much players 'cluster' and the composition of that cluster. Is it a bomber, and 3 nans or a bunch of figs, or a bunch of ints (say camping an enemy base), or a scout and 2 ints (on a miner killing run).
-Track how quickly the team defends against a threat. How long does it take, and how many players, respond to an incoming bomber attack run on a critical base. I think you will find many cases where the response is surprising slow and half hearted.
-Recognize what tactics take coordination
--escorted bomber run
--miner killing run
--galving an enemy base (this usually requires a number of fighters attacking as a group)
--stealth bomber run on an enemy tech base (again requires more than one bomber)
Dog
-First off I think Datamine seems to be on the up and up, they are looking for data from inside a real time strategy game, NOT stats in terms or palyer ranks, etc. Such data is hard to come by. Allegiance can offer it as the code is open thus they have the opportunity to modify the server and get people to play on it, allowing them to freely collect game play data. A server side only change would be best.
-Datamine, you need to realize that while we want to grow, we are a relatively small community and small player base. We have at most one game (of more than 20 people, 10 per side) going most of the time. It may not be worth your effort to mine data from this game.
-We already have our players splintered across multiple public servers. The best way to get players to play on your server is to offer a very low latency server with high reliability (i.e. it is up 99% of the time).
-For your server to be 'well met', you'd need to be listed on our public lobby, for that you need to talk to TigerEye and Pook. It would be good for you to run TAG as well so we continue to get the data we are collecting as well. Getting TAG installed means talking to TigerEye and Pook.
-Providing said server for only a month or two is likely going to be problematic unless we have a good idea of when it will and won't be up. Players need a good reason to play on your server versus other servers.
-I for one would like to see the code modifications you have made (well plan to make) to the server. Recognize that the code is not the easiest to navigate and that you will need a STABLE server. If your server becomes known for being buggy and crashing, people won't play on it.
Data collection wise I think the following (which is only a partial list) would make sense:
-Track time spent in each type of ship. It would be useful to know what percentage of flight time was spent piloting a bomber, or nan (scout equiped with a nanite repair gun), or int, etc.
-Track the amount of resources collected. For Allegiance that means He, so track how many miners are out, and the rate at which He is coming in. A graph of that data is going to be very very telling I think.
-Track how quickly the map is scouted. And then track how well it is rescouted, probed, and KEPT probed.
-Track what tech is bought when
-Track if people are following commander orders (can be a bit challenging to code up but you should be able to get at that at least a little)
-Track how much players 'cluster' and the composition of that cluster. Is it a bomber, and 3 nans or a bunch of figs, or a bunch of ints (say camping an enemy base), or a scout and 2 ints (on a miner killing run).
-Track how quickly the team defends against a threat. How long does it take, and how many players, respond to an incoming bomber attack run on a critical base. I think you will find many cases where the response is surprising slow and half hearted.
-Recognize what tactics take coordination
--escorted bomber run
--miner killing run
--galving an enemy base (this usually requires a number of fighters attacking as a group)
--stealth bomber run on an enemy tech base (again requires more than one bomber)
Dog
DOG PROPERTY LAWS:
2. If it's in my mouth, it's mine.
[unless it tastes bad, then it is yours.]
Raveen, I'm on it!
QUOTE -Datamine, you need to realize that while we want to grow, we are a relatively small community and small player base. We have at most one game (of more than 20 people, 10 per side) going most of the time. It may not be worth your effort to mine data from this game.[/quote]
I haven't seen any other open-source games that can claim so many, though...
QUOTE -We already have our players splintered across multiple public servers. The best way to get players to play on your server is to offer a very low latency server with high reliability (i.e. it is up 99% of the time).
-For your server to be 'well met', you'd need to be listed on our public lobby, for that you need to talk to TigerEye and Pook. It would be good for you to run TAG as well so we continue to get the data we are collecting as well. Getting TAG installed means talking to TigerEye and Pook.[/quote]
The problem with that is that in order to get onto your public lobby, our code needs to be reviewed... which is usually a long process, and we can't afford that kind of slowdown. However, your beta lobby seems to do fine, and I've created an easy way for players to play through the beta lobby or the regular lobby... so that might not be a huge problem.
I was planning on simply not using TAG - making our server essentially "zero consequence".
QUOTE -Providing said server for only a month or two is likely going to be problematic unless we have a good idea of when it will and won't be up. Players need a good reason to play on your server versus other servers.[/quote]
The idea is that when you play on our server, you get a detailed review of how you (and everyone else in the game) performed, using stats that are very different from what is currently collected. I'm hoping that the lure of self-review will be a strong enough lure.
QUOTE -I for one would like to see the code modifications you have made (well plan to make) to the server. Recognize that the code is not the easiest to navigate and that you will need a STABLE server. If your server becomes known for being buggy and crashing, people won't play on it.[/quote]
We're in the middle of making server changes. Once we have something relatively serviceable, we have no problems sharing the code. Although the code may access custom libraries to do ODBC stuff, it should be quite transparent. MFC isn't exactly nice about being compatible with windows.h (what we normally use for ODBC), but it's still clear code.
Actually, the base code for this game is pretty clear. Sometimes I disagree with the way something happens, but it hasn't been confusing at all. Once we beat ODBC into functionality, it should be pretty easy. Really, the only things we're actually changing is adding calls to a database.
QUOTE (Tracking suggestions)[/quote]
Great suggestions, and things that are fairly easy to track!
-Craig
That's the plan. /smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" />Dogbones wrote:QUOTE (Dogbones @ May 16 2007, 08:17 AM) -First off I think Datamine seems to be on the up and up, they are looking for data from inside a real time strategy game, NOT stats in terms or palyer ranks, etc. Such data is hard to come by. Allegiance can offer it as the code is open thus they have the opportunity to modify the server and get people to play on it, allowing them to freely collect game play data. A server side only change would be best.
QUOTE -Datamine, you need to realize that while we want to grow, we are a relatively small community and small player base. We have at most one game (of more than 20 people, 10 per side) going most of the time. It may not be worth your effort to mine data from this game.[/quote]
I haven't seen any other open-source games that can claim so many, though...
QUOTE -We already have our players splintered across multiple public servers. The best way to get players to play on your server is to offer a very low latency server with high reliability (i.e. it is up 99% of the time).
-For your server to be 'well met', you'd need to be listed on our public lobby, for that you need to talk to TigerEye and Pook. It would be good for you to run TAG as well so we continue to get the data we are collecting as well. Getting TAG installed means talking to TigerEye and Pook.[/quote]
The problem with that is that in order to get onto your public lobby, our code needs to be reviewed... which is usually a long process, and we can't afford that kind of slowdown. However, your beta lobby seems to do fine, and I've created an easy way for players to play through the beta lobby or the regular lobby... so that might not be a huge problem.
I was planning on simply not using TAG - making our server essentially "zero consequence".
QUOTE -Providing said server for only a month or two is likely going to be problematic unless we have a good idea of when it will and won't be up. Players need a good reason to play on your server versus other servers.[/quote]
The idea is that when you play on our server, you get a detailed review of how you (and everyone else in the game) performed, using stats that are very different from what is currently collected. I'm hoping that the lure of self-review will be a strong enough lure.
QUOTE -I for one would like to see the code modifications you have made (well plan to make) to the server. Recognize that the code is not the easiest to navigate and that you will need a STABLE server. If your server becomes known for being buggy and crashing, people won't play on it.[/quote]
We're in the middle of making server changes. Once we have something relatively serviceable, we have no problems sharing the code. Although the code may access custom libraries to do ODBC stuff, it should be quite transparent. MFC isn't exactly nice about being compatible with windows.h (what we normally use for ODBC), but it's still clear code.
Actually, the base code for this game is pretty clear. Sometimes I disagree with the way something happens, but it hasn't been confusing at all. Once we beat ODBC into functionality, it should be pretty easy. Really, the only things we're actually changing is adding calls to a database.
QUOTE (Tracking suggestions)[/quote]
Great suggestions, and things that are fairly easy to track!
-Craig
One thing I might suggest,
Players LOVE to respond to something new. (I can't emphasize it any more)
When WeedMan finally got a working version of the LONG LONG awaited Starwars mod people went ape @#(! on it for a few weeks. but the core wasn't entirely balanced and still incomplete and Weedman Abandon the project.
From the looks of the project your going to compile, players will get a unique Statistics review at the end of Game. People love stroking their ego and your statistics and offer them even more ways to do it /smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" />
So, my suggestion is this, organize the grand opening of your Server much like an Event for allegiance. you could coordinate that with Ozzy whome is in charge of Events here. the reason I suggest this is your server is not going to be on the regular server and the more advertising of what you can offer us as players can give us reason to take the time to get over to the Beta Server.
Players LOVE to respond to something new. (I can't emphasize it any more)
When WeedMan finally got a working version of the LONG LONG awaited Starwars mod people went ape @#(! on it for a few weeks. but the core wasn't entirely balanced and still incomplete and Weedman Abandon the project.
From the looks of the project your going to compile, players will get a unique Statistics review at the end of Game. People love stroking their ego and your statistics and offer them even more ways to do it /smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" />
So, my suggestion is this, organize the grand opening of your Server much like an Event for allegiance. you could coordinate that with Ozzy whome is in charge of Events here. the reason I suggest this is your server is not going to be on the regular server and the more advertising of what you can offer us as players can give us reason to take the time to get over to the Beta Server.
That was basically what I was hoping to do. It'll still be a bit before we're ready for something like that, though.Barrager wrote:QUOTE (Barrager @ May 16 2007, 10:26 AM) From the looks of the project your going to compile, players will get a unique Statistics review at the end of Game. People love stroking their ego and your statistics and offer them even more ways to do it /smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" />
So, my suggestion is this, organize the grand opening of your Server much like an Event for allegiance. you could coordinate that with Ozzy whome is in charge of Events here. the reason I suggest this is your server is not going to be on the regular server and the more advertising of what you can offer us as players can give us reason to take the time to get over to the Beta Server.
-Craig
Craig, you might run into a problem with the beta server problem. People don't play on the beta server /blink.gif" style="vertical-align:middle" emoid=":o" border="0" alt="blink.gif" />
If you ask for people to join, you might get a 5v5 game... if you're lucky. The only way to force people over there is through ASGS, which would make Pook want to look at the code first.
So... 5v5 games or look at the code first?
TB
If you ask for people to join, you might get a 5v5 game... if you're lucky. The only way to force people over there is through ASGS, which would make Pook want to look at the code first.
So... 5v5 games or look at the code first?
TB
[18:48] <Imago> dont take me seriouslyspideycw wrote:QUOTE (spideycw @ Nov 28 2008, 02:50 PM) All the retards are contained in one squad mostly (System X)
Well, first, I'm not trying to HIDE the code. People can see it if they want to. It's just that code reviews generally take quite a long time.TheBored wrote:QUOTE (TheBored @ May 16 2007, 12:05 PM) So... 5v5 games or look at the code first?
5v5 games - or 3v3 games - are fine to start with. Then those players will (hopefully) be happy with the metrics we take, and want to play more games. And bring their friends.
That's the approach I'm trying to take, at any rate. Dunno if it will work.
(I should note: I created an easy way for players to play debug.)
-Craig
(Why does the forum log me out every 3 seconds?)
Last edited by craigp on Wed May 16, 2007 8:27 pm, edited 1 time in total.
Well, I hope that can be overcome. My alternatives are not great.Raveen wrote:QUOTE (Raveen @ May 16 2007, 12:38 PM) I fear that you'll find Allegiance players are tricky to move from server to server. No idea why but people are resilient to shifting from place to place even if there's a good reason to.
-Craig