Caha_tavern_b2 waypoints

[TomTom]

Thanks Nova2001. Just to keep you in the loop I have been trying various tweaks but most have only had small or inperceptible effects. Some better camps and tighter defense likely did the most good. I got it down to 2.5:1 ratio in favor of axis in single objective mode (zero initial XP). Campaigns are worse still above 4:1. I am going to try some more radical script tweaks tonight (like doubling the time to own the 4th floor flag to 90 seconds) and see how that goes. Worst map for balancing in my waypointing experience yet one of the best otherwise.

[Nova2001]

I win democracy!

[Nova2001]

Incidentally, I'm going to be incommunicado for the next two weeks. My mysterious, stony silence should not be interpreted as lack of support.

Keep up the good work!

[nedd3h]

i just had a quick game

-Taliesen was standing still at top of 3rd floor stairs
-noticed both teams got messed up when i built the 3rd floor allied ladder fence


when i get some more time i plan to run a test night on my server for this map
more feedback soon!

[TomTom]

OMIGAWD I thought that was just some wood from the local Chinese DIY! I forgot the ladder fence! I'll get on it this week. Thanks Neddie

[nedd3h]

no probs! ...looking forward to your next release

[TomTom]

Well the ladder fence is going to create a problem. It is a problem I have had a number of times with Fritzbot ET. A weakness in Fritzbot is that most actions are either active or inactive globally. This is because Fritzbot uses a global map of nodes that are as it were always active but can be restricted by team. A bot assumes an active action can be navigated, then starts to search for a path. If no path exists it zombies as it tries to find the impossible. Had Fritzbot used a cellular map of nodes the bot would first determine which cells were available then chose from the active actions in those cells.

See what I would propose is;
(1) Add a cell group parameter to the nodes (separate from the group parameter)
(2) Actions would inherit the cell group number of their closenode
(3) Bots when choosing an objective would first check what cells were open to their team at that moment from the cell of its nearest node.
(b) Next the bots would priority-select an active action that was in one of the available cells
(c) Cvops in uniform would check for teamdoors providing access to additional cells
(4) When a node_connect/disconnect happened a cell manager would check if it was one of the links between cells and then update the cell connection data structures as needed.

Ok back to caha_tavern the problem is the combination of the 4th floor axis spawn becoming active before the gen B is blown. If both the stairs wall and the ladder barrier are built then 4th floor axis bots that chose an action on the third floor or vise-versa will zombie until they get killed by the other team if ever. But if the allied bots are spawning on the third floor a stalemate can occur. Without the ladder barrier the 4th floor bots continue to seek ownership of the 3rd floor objectives before proceeding on to the roof.

So what I have to do is either make the ownership of the 4th floor flag destroy the barrier (and prevent it being built) or prevent the axis gaining the 4th floor spawn until gen b is blown (if I can). And this has to apply to humans as well as bots. I'll see what works, stay tuned.
- - -
Update after testing: The latter limitation on 4th floor spawning is feasible by writing a new axis autorespawn manager that choses the highest floor that axis has the flag for except for the fourth floor which requires the "B" generator be blown before the 4th floor spawns are available. As a minor bonus capturing the 4th floor flag early can push the allies respawn to the roof so that axis can get on with owning the 3rd floor objectives. I will likely focus the axis bots on the third floor till the gen is blown. That is so axis bots getting trapped on the fourth floor will be minimized. For simplicity the 3rd floor ladder and barrier will be independant (it is confusing to have to blow the ladder before building the barrier). I have also found a bug in the original flag ownership timers (they can short cycle on recapture) so I have made a basic improvement there. I have still to complete the helper_ents for the third floor logic (I may post a version with a messy aiscript before then). And I am wondering if there is a possible jump exploit over the ladder barrier from the ladder boxes (manually I can't do it but it looks close enough for maybe a scripted jump).

[TomTom]

After an evening of testing I have decided to post the latest work a bit early rather than wait for all the balance testing etc that will take us into next week. With all the changes I have upped the rev to 0.10.

MAP PURISTS will note the following differences from redRum's original:
(1) The cellar wall can't be blown with a grenade
(2) Gen B can't be blown with a satchel (slows axis down for 30+ seconds for balance)
(3) Axis can only spawn on the fourth floor if they capture the flag after gen B is blown
(4) if axis touch the 4th floor flag before gen B is blown Allies are forced to spawn from the roof.
(5) the ladder barrier can be built before the ladder is destroyed
(after all you could build the ladder before destroying the barrier!)
(6) the allied respawn time is equal that of the axis while the object is stolen
(7) there are a few extra message announcements.
and the new cvar Fritzbot_balanceIt can be used to decrease the match time and initial allies respawn time.

[CrapShoot]

Originally Posted by TomTom

and the new cvar Fritzbot_balanceIt can be used to decrease the match time and initial allies respawn time.

nice. i didn't know that could be done.

[TomTom]

Yah the cvar test commands work but little else. Another one of the things not completed in the original WET source.
http://games.chruker.dk/enemy_territory/forum
Debugging_Script_files#G_scriptdebug
FritzBuglist_ET#REQUESTED_KNOWN_ET_BUGS_BE_FIXED_I N_FRITZBOT_ET
Unattended_Testing_Methods#Scripting_Counters