kuroi neko

Yes, definitely! After 500 - 600 XP, a leader will draw enough cards to smoke or down a wingie in a single attack more than 50% of the time. The "distract wingman" skill tips the scales even more, and even a fully defensive wingman ("check six", "natural pilot", "spoil aim" and "sweep tail") will usually go down in the course of the first 3 turns. Even redraws and extra cards are not enough, since a random draw of 4 or 5 cards will rarely be enough to counter a serie of 4 or 5 selected attacks & manoeuvers (useless attacks/manoeuvers and inappropriate counter cards). The more cards drawn at random, the less effective the defense against the same number of attacks/manoeuvers. Besides, on the attacker side, the effect is exactly the opposite: when the wingie attacks, nearly every card drawn counts. Except when the enemy leader is tailed, even manoeuvers will be uselful (an in that case you may want to bully the enemy wingie instead ). Even defensive actions can be used to damage or deplete the hand of the target, so gaining more attack cards make the wingmen proportionally more dangerous. As experience grows, wingmen are both becoming more threatening and more vulnerable, so they appear more and more as the obvious target. As for trying to even the odds a bit, just a few idle thoughts. I am well avare that tweaking the mechanics of a well proven system is very dangerous, so these few remarks should be taken as simple and vague suggestions, not requests or recriminations . Having said that, I guess one could work on both ends of the problem: 1) increase wingmen resilience new skills allowing wingmen to better defend themselves could be added. I was thinking of a defense skill working a bit like "anticipation" (removing a random card from the leader's hand), which would allow the wingman to cancel one card of his choosing during an attack. A lighter version could be a skill working a bit like the "agility" special ability, allowing to turn any card into a strong counter manoeuver, like scissor or vertical roll. Different skills could be used for different manoeuvers, or the efficiency of the manoeuver increased each time you buy the skill again (first time : tight turn, 2nd time: TT & BR, 3rd: TT, BR & SC, 4th: all counters except aces), or the choice of manoeuver left to the wingman as the skill is used. How many times this skill could be used in a given game is debatable, but I think that could end the infamous "manoeuver+IMS2:D" sudden death for wingmen, and lessen the impact of an OTS3:4. To kill a wingie with IMS2:D you would have to have at least a serious other attack or counter card at hand, and be ready to sacrifice at least a valuable defensive card in order to hurt the wingman. As Brian & Dan said, allowing multiple buying of "check six" is a possibility, but I don't like it much because the randomness of the draw will still produce extremely varied results. Either a complete failure or success will depend entirely of a non-controlled random draw, the only strategy will be in buying the skill and computing chances of success. The attacker will have to trust blind luck, while knowing an extra defense card is enough to beat the resilience of the wingman gives more elaborate tactical possibilites (basically, the same kind of choices between attack and defence that occurs while considering attacking a leader). 2) decrease wingmen lethality If new skills allow wingmen to survive longer, it may be necessary to reduce their nuisance potential accordingly. This would make them less tempting preys and allow to come closer to a more balanced tactical choice of target between leader and wingman. The number of w/m attack cards could be reduced by one (possibly adjusting the total attack count against bombers by +1 to compensate). The "sweep tail" skill could be extended to leaders. An element with "sweep tail" both owned by the leader and the wingman would force the opposite wingie to drop 2 cards, which should even the odds nicely even if the nasty wingie survives for the whole game. Wingmen could have to manoeuver against other wingmen to place the nasty 3:xxx attacks. For instance, they could play freely 1 or 2 burst cards against other wingies, and the 3 burst cards only after gaining at least a manoeuver point with "manoeuver" or "half loop". This would tone down the lethality without ruling out a nasty OTS3:4 from an inspired wingman to an unfortunate one. This would be especially nice in low XP duels, where a random OTS3:4 from a one attack card wingie has an unfair devastating effect on the opposite one defense card wingman. [ October 13, 2005, 11:51 AM: Message edited by: kuroi neko ]

Thank you, Brian. By the way, your last change to the card drawing logic seems to work like a charm. Now I get one or two IMS2:D per game, exceptionally 3 (which is ok since high level planes are drawing enough cards per turn to dry the deck more than 2 times in a 6 turns game). The annoying "IMS2:D avalanche" never reappeared. Congratulations . Sending this "mortal wingie" bug down in flames will get rid of the last serious implementation problem, as far as I'm concerned. After that, a bit more variety for high-end computer planes and pilots would be nice. Right now I only use these games for regaining wingmen EDs, so this is no such big deal for me. However, I found the AI pilots level a bit unnerving while I was trying to unlock the top Japanese planes. Preventing players from racking up kills against bots is one thing, but killing them just for trying to see all available planes may be overdoing things a bit . I know tweaking AI routines is a tricky business, but frankly the super-über aces are showing the limits of computer logic: with the dozen cards they expend each turn, they should wipe the floor with the planes players can oppose them. However, since they fail to go for wingmen first and discard far too much of their cards, they can be beaten (not easily, a simple bad initial draw and the player has to run for his life, but usually the bot wingie buys it and you end up with a smoking leader) while they should simply not be .

Sorry to insist, but could you please have a look at the reason why wingies practically never regain their ED? As it goes, I'm stuck playing silly escort missions against bots to regain my wingie EDs because my demo wingies with their so cool medals are dying as soon as I play against humans. Bots kill them too, but slowlier (since the 8-skills & 7 redraws AI did no grasp the fact that killing wingie is key to success). I lost another one of them yesterday. Besides, playing escort around 600-1000 XP wears pilots down at a preposterous rate and result invariably in 50 XP loss per mission, my Japanese are all flying with between -2 and -4 structure points. However, I found these are the only missions which allow wingies to score a kill once in a while without being killed themselves too often. This obvious bug is simply grounding my high-level pilots and forcing me to be bullied around for hours by über bots. I'm starting to get annoyed quite a bit. It should be only a couple of lines of code to look at and fix, not a major design change. Please save the wingies .

Cool feature!

I'm honoured

About purely graphical inaccuracies : Hurricane I - typical early war cammo would be green and brown, with lower surfaces painted half black, half white (was intended for easier identification from the ground(!)) Spitfire I - had 8 7.7mm MGs, no cannons (except the Ib, but only 30 or so were built, and only one squadron was equiped with this variant during BoB). - the typical camo scheme would be green and brown with ligh blue-grey lower surfaces, the green and grey is more suited to 1943 and beyond (perfect for the spit IX ) spitfire XIV - the color scheme is too irregular - both green and brown should be darker 109B - typical cammo would be overall medium grey with Spanish markings (black X-shaped crosses on a white rectangle) - square wingtips - completely different cowling - tail struts 109E the model looks like a 109F to me - wingtips should be square - typical cammo (France and early battle of Britain) is sky blue up to the middle of fuselage and two tones of dark green on the upper fuselage and wings - spinner should be more cone-shaped with a blunt tip - side air intake should be smaller and more square - tail struts still missing Stuka Green should be a bit darker . 109F good model except the wrong spinner shape cammo could be African orange-yellow 109G - missing the supercharger on the cowling - wrong spinner shape 109K - camo should be closer to the one of the 109G, may be light grey and green mottling on the fuselage - wrong cowling, missing superchager and 12.7 mm guns - the usual wrong spinner - most 109Ks had a reshaped canopy with only 3 panels P40E - strange grey cammo, not sure but I think the same pattern in medium green and sandy brown would be more representative. [ October 11, 2005, 11:06 PM: Message edited by: kuroi neko ]

Lots of questions and affirmations in there About historicity, it is clearly not the main focus of the game. However, having being a passionate IL2 player and done my share of online flying, forum bashing and virtual squadron living, I found this little "simplistic" DiF a truly excellent abstraction of aerial combat. The bases of energy management are surprisingly well implemented by the cards system: you must use your cards wisely, waiting for an opening and trading energy for fire solutions. Spending a turn or two building up your hand simulates pretty well the initial tactical manoeuvering, trying to get into a better position without engaging. Also the need to plan ahead is a pretty good abstraction of the situational awareness, especially for games with 2 elements. You can decide to focus on the plane you're engaged with but you'd better not forget about the other planes around. Generally speaking you need to keep an energy reserve after a gun pass or you'll be a sitting duck, but the global situation may call for exceptions: you may want to down a very good pilot (i.e. with lots of skills) if you get the chance, at the cost of a possible revenge from his less experimented wingman, engage fighters with no real means to harm them just to prevent them from having a go at your bombers, etc. The cards are very well chosen, they manage to capture a stylized representation of the basic options in fighter combat. For instance, the "scissors" are indeed a defensive manoeuver that can allow to regain initiative only if an attacking plane has been willing to sacrifice energy to get on your tail. With enough energy (i.e. a strong remaining defensive hand) the attacker will escape the scissors and continue to dictate his conditions to the defender. Of course the name of the cards are to be also taken as abstractions. For instance "barrel roll" and "vertical roll" are abstractions of horizontal and vertical offensive/defensive manoeuvers. And lastly, the pilots skills are an excellent way to implement the "it's the pilot, not the plane" saying. You can offset planes characteristics by having an ace at the controls, but even an ace will have to sweat for downing a truly superior plane. Here again I find the trade-off between tedious level of detail and playable abstraction very well tought out. A "raw" plane represents what a wet behind the ears pilot can make of it. Adding skills represents what is learned in combat situations. After all it took me dozens of hours of virtual flying to become an average shooter in IL2, and frankly I'm grateful this game offers me a chance to experience what being a marksman fells like just by accumulating enough experience to buy the skill . This game is not IL2, where people argued for months (or even years!) over the undermodeling of the 190A8, the loss of ShVak rate of fire due to improper propeller synchro or the 109G-6 vs LA5-F power curve . True, the choice of planes is rather debatable, as are the color schemes. One could bicker for hours about the 109B and K being painted in African camos, the spit I looking like a spit IX with his pair of strictly a-historical cannons, the zero limited power rating not being enough to simulate his vertical agility, the P40E being over-modelled in comparison of how she fared historically against Japanese planes of the era, and so on. But I think engaging in such debates is missing the point of the game entirely. The strength of the system lies in its simplicity, and even though the graphics could be enhanced toward more realism, the core of the game seems to be a very well proven design and has to remain streamlined to work. Is that game realistic? Not in many small ways, but broadly speaking, yes. But the question I'd rather ask is: is it it fun? I'll let you guess what my answer to that one would be . Concerning the spitfire V / hurricane II comparison, dont forget the wingmen capabilities : 2 and 3 cards but the spit is better on defense while the hurri has an very good attack value for planes of this level. The firepower of the hurri makes it dangerous in a hit & run situation or against bombers, but the poor wingman defense will get him into trouble facing a pair of more nimble and/or powerful planes. About special rules for MK108 (or even MK103!) cannons, the fighter vs bombers system already takes the global offense/defense capability of fighters into account by adding the attack and defense cards of the wingmen. I find it pretty good and sufficient. The only plane with the MK108 is currently the 109K, and it is reflected in the wingman offensive capability. It causes a real disadvantage when attacking bombers without a wingman, which simulates the "mass effect" of fighter vs bomber attacks, yet again in an abstract but pretty consistent way. About campaigns (however ), I completely agree with you. A pilot(s) career simulation could be a lot of fun, although that would have to be a solitaire kind of game, which does not seem to meet the main design choices. [ October 11, 2005, 09:26 PM: Message edited by: kuroi neko ]

Ah yes, I had the same problem. This is because for some obscure reason the regular variant of the Arial font was lost after some windows auto-update. To solve this you just have to re-install any regular arial font, you can find it on the web rather easily. And there is no danger of getting a virus, a TTF file is pure data, it wont bite or anything.

Well the manual says surviving the mission is enough to get the 20% chance, but there must be another "hidden" condition, like maybe not having being damaged as you said. It would be nice to have this clarified.

lol ok, that part about random generator choice was just a private joke About threading issues, I suspect there is another quirk in the card drawing interface. In the discard phase, if you click the "discard" or "next turn" button just at the time the timer runs out, you may end up with all your cards gone except the ones you just redrew. It happened to me only once (during a campaign turn with Horatius). I had a 5 cards hand and 2 power points, discarded one, then typed something in the chat and forgot about the timer, so seeing that I was runing out of time I clicked in a hurry, but just a second too late. If I remember correctly, what happened was that the next confirmation screen did show up very briefly, showing only the 2 new cards, then my turn ended. At the start of the next turn for this element, I saw I had only the two new cards in my hand, all the 4 old ones were gone for good. It looks like the timeout event fell just in the middle of the redraw procedure and messed up the cards list. Just an educated guess, but it could come from the timeout being handled in some kind of callback or service thread which would access the global player's card list (by calling the "next phase" activation or something like that) already being modified by the main user interface thread. To avoid that, I suppose some shared data access protection with the user interface thread could be added (mutex or equivalent), or the event could be re-routed as a message to be handled directly by the thread already in charge of the rest of the user commands.

more on "wigie cant regain ED" topic: Yesterday I finally got an ED back for one of my old wingies, and it may be a coincidence but he happened to shoot down a plane on that mission (which, as it was already said, becomes more and more difficult as experience increases). My hunch is that shooting down planes is the only way to regain ED for wingies in the current implementation. I wonder if some kind of automated test could shed some more light on the problem, like the server code creating fake player elements, playing them for a while and doing statistics on the ED recovery?

Excellent news Frankly I'm rather impressed with the rate at which you adress issues! Will I dare to ask which kind of random generator you used in this new version?

Well my intend was certainly not to discourage people from buying this excellent game . Frankly (and I owe no share in DVG ) the designers are extremely open to suggestions. In less than 2 months they did a considerable number of changes, most of them based on direct user feedback. I am quite confident that the system has a great lot of potential for evolution, and I reckon buying the game in its current state, despites a few quirks, is the best way to support the team and get a chance to see all these very promising extensions one day.

We were discussing the matter with StalinsO and a few others, and we thought maybe you could add a different matchup system. Let's say you and your oponent agree on a given level (number of XPs or resulting points), then both go to the selection screen, choose a nation, a plane type, buy skills, and play the game at whatever level both players agreed on. Of course the pilots would be unranked, no experience gained and no leaderboard kills recorded. But that could relieve people from the frustration of having to play 10 gladiator/nate/109B/P35 missions before getting their hands on a "decent" plane (or rather 30 or 40 missions, since pilots must rotate due to fatigue, and thats not counting the huge number of games needed if you like to experiment with all 4 nations).

You can find it on various abandonware sites. Still quite fun after 10 years

Well this seems a very good way of displaying all the informations. Reminds me a bit of the icons in "over the Reich" and "achtung! Spitfire". I always liked the simple and readable design of the squadron roster in these games. An intermediate possibility would be simply to create this log file I talked about, very easy to implement in comparison with new graphic stuff (although some modification to the protocol would still be needed, but that will have to be done either ways). It may be enough to see what really happened and get a link to what is recorded on the server (a simple timestamp or game identifier should do the trick).

Ok, I did quite an extensive testing myself and could not pinpoint a fault. Especially the "wingman draws a new card each time I play one against him" case. Either this problem sorted itself out in new software releases or it was just a figment of my initial frustration against the über bots . After all my ranting, I am ashamed to admit I may have been fooled by the amazing amount of skills from the über bots at the demo time, combined with the "exotic" card deck implementation . However, allowing players to know the exact reason of each AI redraw would put an end to all suspicions. A small log area in the game window could be added, with inforamtions like "pilot ppp draws xxx cards at start of turn", "pilot ppp uses skill sss to draw a card", "pilot ppp discards xxx cards to follow altitude change", "leader lll drops a random card due to wingman www skill", etc. Also, a detailed log file of the last mission could be created locally to allow players to get rid of any doubt about what actually happened during the game. As an added benefit, this file could also be used as mission report (instead of the current screenshot required for tournament games).

LOL even I dont spend enough time online to satisfy all the potential bloodthirsty players playing with my poor little "smiling face" guy. Seriously, a semi-random massively multiplayer campaign could be a LOT of fun. Here again I'm drawing from the excellent IL2 flight simulation community: 2 or 3 years ago there was a system called Virtual Eastern Front (VEF) that allowed to play semi-historical IL2 games. The game was open to any individual player, provided he had registered into a given squadron for a given country. It was possible to have 8 vs 8 missions with a mixed bag of squadron members, but of course people soon started to form "serious" squadrons flying together all the time. The pace of the game was "simulated real time", i.e. about one game week for a real life day, which allowed to cover all WWII in about a year. The war was broken down into separate campaign areas (famous eastern front battles like Moscow, Stalingrad, Kuban, etc.), each with a given set of available planes, ground targets and strategic objectives (i.e. tank busting for Kursk, air superiority over Stalingrad, transportation interdiction around Moscow, etc.). To start a game, players gathered in a VEF dedicated section of the game lobby, and as soon as there was enough people to play, the host would go to a web page and generate a new mission (which is basically how the current DiF campaign system works, except that the DiF campaigns are limited to 1v1). The host would choose general type of mission: number of players, area of attack, main goal (air superiority, ground attack, etc.) and generate a scenario. The mission was then played and the results sent back to the VEF server. The global situation was updated (slightly) after each game: global victory points, available planes, ground target locations, airfield status. Even the front line could move. Since the changes were very gradual, dozens of simultaneous games could be played without shifting the global situation too abruptly. Only after a while you would see the front moving or notice a change in the global choice of targets and mission types. For instance, you could find yourself forced to start from far away areas after all your front-line airfields had been repeatedly attacked, notice that a few flak bateries were gone from your usual flight path, etc. The strategic situation would not change whatever the players achieved, but tactical successes or defeats could reshape the front line or change the pace of historical events for a while: you would not win the battle of Stalingrad by shooting down hundreds of Russian planes or destroying hundreds of Russian targets, but you could somewhat enlarge the pocket and delay the defeat by a couple of weeks (or hasten it if your Luftwaffe guys were regularly trashed by the VVS ). I guess it would be possible to design something similar, adapted to the DiF campaign system. Just a few ideas that I found interresting at the time: Even though DiF campaign system does divide a territory into fixed-size areas, maps with half a dozen zones could be enough to simulate front line: there could be friendly and enemy zones plus some "neutral" front line areas with a given percentage held by each side. The current objective per zone system seems to fit very well into the picture: the choice of objectives would reflect the global situation evolution, and the result of each mission would in turn influence the available objectives. By choosing objectives the players could try to shift the situation toward a global goal (concentrating on a given type of objective across several missions). Or grab a tactical oportunity (a juicy truck convoy wandering near the front lines) Or simply pick one of the available objectives at random . Starting airfield location (influenced by airfield attacks) could be simulated by the same range system as the current DiF campaigns, but in order to add flexibility the range could have an effect on pilot fatigue rather than be a simple limitation of reachable map areas: flying over far away zones would simply cause more fatigue. The flak density could also have an impact on fatigue or planes wear and tear, thus adding another tactical parameter (some flak suppression missions could allow to clear a zone from some of its AAA defenses). Squadrons could run out of a certain model of plane and be forced to switch to another type while waiting for replacements. The available planes could be limited not only by the number of shot down planes but also damaged ones or even the wear and tear tied to the number of sorties per day. People would have to be careful about getting smoked in a scenario with few planes available, or would have to make each mission a success to offset the attrition rate of their planes. For instance, at the start of the war, Luftwaffe usually had a smaller pool of available planes, but the VVS had vast quantities of second grade planes (I-15 and I-16) an a precious handful of modern fighters (LaGGs, MiG and Yak). In such a situation, Russian players would hope to get a mission with top notch fighters but would have to fly them carefully, while the usual ratas could be flown somewhat recklessly (provided the pilots got away with it!). Fighter-bomber missions could be simulated with a slight extension of the current game system, like the possibility to carry various external armaments at the cost of reduced power and/or manoeuverability, plus a new "jettison all ordnance" button useable at the start of each turn . 2v1 escort missions would also fit very well into the current game system, the campaign would simply add limitations to the number and type of available bombers. Pilots bailed out over enemy territory would get a chance to escape capture, or be held prisoner with a slight chance to escape over time, or be returned to duty if the front line shifted beyond the location they had been shot down over. In terms of game it means they would be prevented from joining games for the time of their "captivity" until the end of the campaign or their liberation/escape. Well there is a lot of room for imagination there, but after all the "hornet leader" game has already been designed by a certain Dan Verssen, so I guess the idea could be worth a thought. [ October 09, 2005, 03:58 AM: Message edited by: kuroi neko ]

I agree on the "wingie must go first" part. I think the increased nuisance caused by wingmen coupled with decreasing relative survivability designates them as first-choice victims after they collected 800 XP or so of permanent skills. At this level, even a fully defensive wingman (check six, natural pilot, sweep tail, spoil aim) should have a very short life expectancy. This is one of the major areas where the bots fail to play properly, by the way. The AIs are going for the leader far too often, that is one of the main reasons why the über aces are still "winable" despite all their skills and redraws. Since the number of structure points does not rise as quickly as the potential damage per turn, the poor wingies become at the same time more threatening and easier to kill. The two effects add up in favor of getting them out of the way first chance . And the general nuisance of the most powerful wingman skills (distract leader, leader advice, extra energy and so on) add yet other reasons to go after them first. I also agree on the wingmen not regaining their escape death. I take the word of Dan that the intended rule is 30% to regain the ED after a mission, but the implementation must have lost its way somewhere. As many other "frequent flyers", I have wingies who never got their ED back after more than a dozen missions. The probability to fail to get the ED back would be something like 2% after 12 games. And this is not an isolated case. Besides, regaining ED seems to work for leaders. But maybe it is just because of the second possibility (after scoring a kill), since leaders usually kill a lot more than wingies.

Yes, that is a very good idea. Since the game is heavily based on client-server architecture, why not take advantage of this to design a kind of "oponent finder" web page? It should be possible to check which people are connected, and to display all comparable pilots for a given player. A simple click on the desired pilot could even trigger a game request directly on the opponent's client side (well of course this should include some sort of ignore/away system).

More experiments: I got a game where the bombing objectives were all battleships: 3/5, 3/5 and 4/7, plus the usual air superiority missions over malaysia. I would say such a setup is impossible to win for the Japanese against a human player. With such a setup, the obvious wining strategy to put the best elements over Singapore, since even a draw on the bombing missions can give more points to the UK than a win in air superiority. Basically the kates cannot hope for better than a draw, and against a competent player there is no way they can make it to the target with 3 bomb points left (damaging both or downing one is enough to go down to 2 bomb points for a UK win). So that is practically a guaranteed 6 points for the UK. Now the bettys can maybe win once (yet again damaging both or downing one is enough to get a draw), but except with very favorable starting alts there is no way the Japanese fighters can prevent the 2 British elements from hacking the bombers to pieces. That will most likely end in a draw, and even so the Japanese will have a though fight. As for the last objective, it's again a draw at best, and frankly the bettys will be in so poor shape that the most probable result is a loss (downing one is enough to get a UK win). That makes 8 more points for UK. Even assuming the Japanese wins all air superiority missions, the likely score will be 14 to 6. I managed to get a 6 to 6 result only because the AI stuck to my fighters instead of going for the bombers, which is the obvious thing to do for a human player. The AI won 6 points against the kates in turn one, the Hurri II element started co-alt with the bombers and after two turns had one kate down and the other smoking . I have no gripe against this, but I wonder if we will play enough games in the tournament to even the intrinsic randomness of this setup.

Very well put, Brian As for the actual XP used, I'm rather surprised because during the demo period it seemed to me the bots were a lot tougher for a 2000 XP element than for a "true" 100 XP element. After full release I only had quite experienced pilots (maybe 300-400 XP for the "weakest" pair) so I could not really test the difference. It may well be a subjective extrapolation of the impressions I had from the demo.

Ah cool, thanks . Sorry for the endless ranting. I think this will be a noticeable improvement.

Hehe I do take interest . As I see it, this system gives an advantage to pilots with permanent skills. This in turn gives an advantage to frequent players, since it is psychologicaly easier to spare the XPs if you play 5 or 6 games in a row and get the "reward" in a single gaming session than waiting a whole week to see your pilot evolve if you play one game per day. Another downside of the current system is the choice of bots, which is still based on the XP total earned in the demo. A 2000 XP pilot with a "post-demo" total experience of 500 XP or so will most of the time face monstruous bots (with about 6 or 7 permanent skills and tons of redraws), while a pilot created after the demo with the same "real" total will face "manageable" AI pilots. This is mainly annoying for 2v2 games with only 3 human players: the guest bot is often so horrible it destroys the whole game balance. But after all this is just a demo problem, it should disappear with time. As for keeping the current system, I would say it rewards players who care about their pilots survival: buying cards does more for the offensive than the defensive, since the rarest and most expensive cards are basically attacks, and buying cards for defense is really a strange strategy, given the very limited defensive use of even an "ace pilot" or "vertical roll" over the devastating effect of an IMS2:D or OTS3:4 in one game.

I get a very minor graphical glitch sometimes at the last turn (for games involving bombers with less than 6 turns). The rectangular area where the turn and element counters are displayed turns plain blue (like the sky background) at the very begining of element #1's wingman turn and remains so until the end of the game. As far as I noticed, this happened only when element#1 was played by a bot. Win XP SP2, (venerable) GeForce3 Ti200, Athlon 2800+, 1Gb ram