What we end up doing will be whatever works best, and I will depend on you guys to tell me what's best. I'm open to try anything. But we need to start getting some real data on whether these tactics are currently broken.

Sing, I'm going to go back to an earlier version, prior to these timing changes, then add in code to measure actual delays for all of the synched actions, plus measure real pings and event response times. Then with that data I'll be in a much better position to reproduce the expected timings. That'll take me a couple of days to put together, then I'll put it up on twgs.classictw.com and call for players to come in and run scripts as I gather data.

I really don't think they are. We did some timing and it
looks pretty spot-on as it stands. If I thought we were
way off, I would've said something much earlier.



I've already got a lot of timing data. What exactly do
you need timed?

Well, if you really think it's in the ballpark, then I guess I don't need any. I thought you said that the timing needs to be very tight, and I really haven't had any feedback yet telling me that the current settings are adequate. I figured the only way to get an accurate measure of what the timings were would be to go in and measure them internally. But obviously I won't waste my time doing that if there's no reason to do it.

If you think the timings are up to it, maybe we should see about having that SG tribute tourney on twgs.classictw.com under this version, to really shake it out. That has been suggested, but I didn't think we were anywhere near ready for that.

Not to step on Bud's toes, he can still go forward with his SG Tribute Tourney. I just think it would be great to run the beta test game on twgs.classic.tw on one of SG's edits.

And I think it would mean a LOT to this community if JP would do us the honor of playing along side us in the beta test game. What do you say JP? Come play with us!

I looked into the game, and game L says Sub Zero but the the ships and planets are stock. Isn't it sort of hard to test some scripts in an edit where planets that advance that slow?

With a 150 / 1mbit setting, I think we'll be pretty close. A little
tweaking might be in order at some point, but I think those
settings are good for a wide beta.

I turned Gold off on the SubZero edit to limit the scope of the tests. I'll expand the test to Gold features soon.

Yes, JP asked me which of the edits to run, I told he he'd be better off asking you or Sing. I've never played SG's edits but I have seen a SubZero and I know the one running there now is simply a stock game with 30k sectors and unlimited turns.

He told me these are SG's edits saved on that server now:

1) ClassicTW Stock
2) Gridwars
3) 2010 Haunted Halloween
4) Ice
5) SubZero
6) Parrots Place
7) The Social
The Classic Unlimited
9) StarTrek Role Playing

I suggest someone who's familiar with these edits to tell JP which one to run the beta test game on.

Thanks guys!

SG didn't make StarTrek, Subzero, or ICE.

Subzero is a good aggressive edit, but you
will need to rebang it every few days to
keep it fresh.

You should also try a Pirates and a MOO2.

I've noticed that the port/planet buydown has very little, if any delay. When using mombots buydown it actually uses a lot of my cpu, so I'm sure it's lagging the server. I'd suggest putting in some sort of delay there. I haven't been able to test torps or pdrops yet, but the timings seem pretty good other than that so far.

Right now the SubZero is running at the fastest settings

I've just set the Sandbox game to default on movement delays and emulating the speed of a 56k modem.

Looking at them side by side - running WSST on the SubZero and World Trade on Sandbox you can see the difference. One is a blurr and one you can make things out.

For your review, try them both on for size side by side and see how the 56k feels. Not bad for a low turn TL game, I like it!

Good Job JP!

While we are testing only, let's adapt a "catch-N-release" policy. There's no need to go killing other beta testers. We need all the traffic we can get.

I take it you got killed too

Yeah, but I expect to be killed. I don't play unlims, I run trading scripts to be someone else's target.

I'm only fodder for the attack scripts, catch me if you can!

I was hearing complaints from others, that's why I've asked for the "catch-N-release" policy. We're NOT playing this game, we're only testing.

When JP is ready to open one of the other games, with gold edits we'll play that one for real.

Ok, over the last few days I went through and calibrated all of the internal timings to a set of default values and I'd like to see how those feel. Basically, what I did was benchmark every action on the game, then set timings so that looping on any action would draw no more than 5% CPU load on my test machine (a 64 bit 2.53 Ghz Intel Core 2 Duo). I've set the test game L on the beta server to these timings, so have a look at them and let me know how it feels.

Once the pace is locked in, future improvements in CPU and hardware speed will tend to drive that CPU load down rather than speed the game up. It'll be nice when the maximum load for a player is < 1% :)