Tobold's Blog
Saturday, May 18, 2019
 
Win rate maths

I get the feeling that I don't understand the mathematics of win rates in World of Tanks. My own win rate fluctuates between 48% and 49%, which is what one would expect for an average player. Slightly below 50%, because sometimes there are draws, and then nobody wins. So the overall average between all players is somewhere in that region. What still puzzles me is to come up with some formula that relates player performance to his win rate. How good do you need to be to reach 50%, 55%, or 60%? How high can you get?

A starting point is probably looking at how much damage you deal in a battle, divided by your own hit points. That ratio is probably very close to 1.0 for an average player. A player with a ratio of 0, basically going afk at the start of every battle, will presumably have a win rate lower than 48%, because in every battle his team has one tank less. But I don't know what that win rate would be, statistically speaking, because I find it hard to estimate how much one tank less would lower the chance to win.

At the other extreme, a player with a ratio of 15 is the theoretical maximum, which in practice will be impossible. But that player obviously would have a win rate of 100%, because he kills all the enemy tanks in every battle. But how does that scale down? A player with a ratio of 5.0 would already be very, very good. And his win rate would be higher than 50%, but I can't estimate how much. The best approximation I have is watching Quickybaby on Twitch, because doing 5 times his hit points in damage seems to be what he is able to achieve quite often, and his win rate is 60%.

What confuses the statistics is that some people play in groups, as platoons or in events for clans. With a platoon or clan taking up more than 1 slot of the 15 tanks per side, they can have a larger impact, and thus win more consistently. But for random battles it appears that between 40% and 60% win rate is about the realistic range that players can have. And I find it very, very hard to move my own win rate even by just 1%. And obviously it is getting harder all the time, because I now have 9.4k battles, so I would need to win a lot to make up for the past. No wonder some people make "reroll" accounts.

Labels:


Wednesday, May 15, 2019
 
Clarke’s Third Law

Clarke’s third law states that “any sufficiently advanced technology is indistinguishable from magic”. If people know this law at all, they relate it to fiction, especially science fiction. But more and more I come to think of it also as being able to explain some events in the real world.

Climate science is so complicated, so based on statistics and models that few people understand, and so hard to demonstrate with experiments, that it fulfills Clarke’s criterion of being sufficiently advanced to be indistinguishable from magic. In particular it is indistinguishable from the miracles and “acts of god” of religions. A catastrophic climate event, like a hurricane, is indistinguishable from an event caused by the “wrath of god”. Even a climate scientist can’t reliably predict those events, current science only tells us to expect more of these events as a result of global warming.

This gives climate science an aspect of doomsaying magic, and of religion: “Repent, sinners! Abandon your acts of gratuitous consumption of meat and fossil fuel, and the world might still be saved!” Some of the people talking about climate change end up being indistinguishable from high priests of some strange religious cult. And thus we shouldn’t be surprised that some people simply choose not to believe those climate change priests. After all, breaking several of the ten commandments doesn’t get people struck by lightning. And people quite like their gratuitous consumption, their steaks, their big cars, their airline travel holidays. Why give all of that up for a badly defined chance of less future doom? Especially if you don’t live in some oceanfront property or otherwise low ground, keeping up your consumption appears to be the safer bet.

As a scientist I am pretty convinced that climate change is real, that it is increasing the likelihood of catastrophic events, and that it may lead to rising ocean levels which might displace or kill a lot of people. But I am also still flying, still driving, and still eating meat. The economist in me calls this the tragedy of the commons. As I can’t even convince myself to lead the life of a minimalist, I also don’t believe it will be possible to convince everybody of the necessity of doing so. The scenario in which humanity manages to limit global warming to 2 degrees Celsius is extremely unlikely, in my opinion. While the doomsaying high-priests of climate change have my sympathy, I don’t plan to practice that particular religion. I would however advise you to quickly sell that oceanfront property and move further inland, if you can.

Monday, May 13, 2019
 
Rage of Demons: Session 14

In the previous session the group started to collect the materials needed by the drow archmage Vizeran to banish the demon lords back into the abyss. In this session they continued this "shopping list", moving into an area of the Underdark called The Labyrinth.

On the way there they encountered a large cave with a strange tower made completely out of adamantium in it. It became quickly obvious that the tower was magic, and while there were a door at the bottom and a trap door on the roof, there was no way to open those. Finally Mog'burz teleported inside through the arrow slits with his Misty Steps spell. However the tower wasn't empty, so he got attacked by a Boneclaw, which is basically an undead monster that is created when some mage tries to attain lichdom and horribly fails. Arkoy then teleported the other group members into the tower, although they ended up on the wrong floor due to the randomness of the teleport spell. The group killed the Boneclaw, and found the diary of the mage describing how he wanted to become a lich. The diary also told them that the tower was a Daern's Instant Fortress magic item, and gave them the activation words so they could open the doors, and turn the tower into a small cube and back into the tower. I did drastically reduce the damage the tower does when he grows from the cube, because otherwise nobody uses the thing as a building, but only uses it as a reusable fireball kind of effect.

Then the group came to the entrance of The Labyrinth, decorated with two piles of heads, and scrawlings in blood praising Baphomet. Surina used the Stonespeaker Crystal to try and talk to the freshest head with Speak with Dead; but as the crystal had become infused in Gravenhollow with the power to grant visions, she learned about the eternal war between Baphomet and his minotaur followers against Yeenoghu and his gnoll followers. She also learned that a goristro servant of Baphoment was hunting in The Labyrinth, at the same time as Yeenoghu himself with some gnolls.

Now the group could have staged the two hunting parties meeting each other, but instead they went for the simpler plan to attack the two parties separately. The goristo and the minotaurs was a tough fight, but they killed him and got his heart for the ritual. Yeenoghu with his gnolls was much, much harder, especially with him advantage on saves against spells, innate spellcasting, legendary resistance (which is automatic success on saving throws), and legendary actions (actions done outside his turn). Even with the help of amulets of protection from demons the group just managed to wound him with two arrows, grab the arrows with his blood on it (which they needed for the ritual), and flee by teleporting away. Yeah, high level groups teleport a lot.

Using a Find the Path spell they then found the gallery of angels, where eight angels were petrified for eternity for sins they had committed. This was a fun little encounter, because the angels can't be liberated, but afflict anybody touching them with various effects. Besides Mog'burz getting temporarily charmed and hitting Arkoy, nothing serious affected the group; although Quarion was much spooked by an angel telepathically asking him repeatedly to kill him (which is impossible) and not shutting up about it (fortunately the effect only has the same range as a voice). They managed to get a petrified feather from each angel, and thus now have all the ingredients for their ritual. As a bonus they could ask the angels a question, twice, and asked the same question twice: Will Vizeran betray them? They got twice the same answer, that it isn't them that Vizeran wants to betray or harm.

After bringing all the ingredients to Vizeran, the group must now wait for the Dark Heart to be assembled by the drow archmage. But in the meantime they got a call for help from the ruler of Neverlight Grove, Basidia: Zuggtmoy has finally arrived in the Underdark region known as Araumycos, which is basically a continent-sized fungal lifeform. There she plans to hold a "wedding" between her and Araumycos, which would give her control over him and the whole area. The group is called to prevent that, which is what we plan to play in the next session, The Fetid Wedding.

Labels:


Saturday, May 11, 2019
 
Progress vs. playing what you like

This month in World of Tanks there is a Top of the Tree event that speeds up progress along the British medium tank line. The only problem with that is that it turns out that I don't especially like the British medium tank line, particularly not their low alpha guns and bad armor. I got to the tier VIII tank, the Centurion, but end up frequently in tier X games against very good players, so I don't do very well.

The obvious solution is to play the tanks I like more, and in which I do better. But that means pretty much ignoring the progress part of the game, and *not* climbing up tech trees. Which is certainly a completely viable idea; it just makes the whole experience point, tech tree and reward system pretty much pointless, except for crew xp.

I think I will do some of that, in combination with another idea: Try to find out which tanks I think I would like, and do their tech trees. That shouldn't be so hard, as I probably don't want many high tier tanks. I have learned a lot about how to play the tactics of World of Tanks in the last months, but I can't beat the reaction time and aiming skills of the tier X players half my age. Unless I play artillery, which for some reason I don't enjoy very much, I better keep out of tier X games, which means playing tanks only up to tier VII.

Labels:


Friday, May 10, 2019
 
Missions in World of Tanks

When I restarted World of Tanks last Christmas, I was very excited about the game now having these missions and campaigns of “quests” to do. I recently finished the first campaign, but I am already a lot less enthusiastic about them. Missions don’t really seem to add anything to the game. While in other games, let’s say World of Warcraft, the quests made you explore places you otherwise would have missed, in World of Tanks missions just ask you to fulfill the role of your tank. That meant the first campaign was somewhat helpful as a tutorial, but once you know what each tank type is supposed to do, missions became just the equivalent of “play well”.

There isn’t really a good reason *not* to have activated a campaign mission for whatever tank type you are currently playing. If you have a good game, chances are that you will fulfill your mission at the same time, and get some extra reward. The non-campaign missions, like the new premium account daily missions or the event missions, are even more along the lines of just play and get extra rewards. So in the end of course the extra rewards are nice, but the campaigns and missions didn’t actually change anything in the way you are playing. Except that to finish the campaign I had to play artillery for two days to get through the 15 artillery missions, which I otherwise might not have done. Curiosly I am better at playing arty than some other tank types, but I don’t enjoy it as much.

I was thinking that missions which would give people goals in battle that have nothing to do with the regular game would probably be a bad idea. You wouldn’t want your team to lose because some players were concentrated on collecting easter eggs instead of fighting. So that probably limits what the developers can do with a mission system in the game. But the result is that I now just more or less ignore the campaign and missions, and just cash in on the extra rewards which I earn without even trying.

Labels:


Tuesday, April 30, 2019
 
Aging PvP multiplayer games

In my ongoing quest to improve my skill and performance in World of Tanks, I am watching good players on Twitch streaming their games. Streamed games are far better suited to learn from, as they aren't the carefully handpicked exceptional replays you find on YouTube. So I was watching QuickyBaby's Centurion AX Tech Tree Showcase, in which he is playing a line of tanks from tier 1 to tier 10. And at tier 1 he encountered a player who had created an account to play over 11,000 games with the same tier 1 tank.

Why does somebody play all his games in a tier I tank? Because that way he can achieve a 70% win rate. Not only does playing the same tank over and over result in you probably having better skills than the actual new players in tier I; you will also have a crew with far better skills, as well as the best equipment on the tank. Having better concealment and view range you can basically be mostly invisible, killing "noobs" without them even knowing what hit them. In World of Tanks that is known as seal clubbing. And it obviously leads to a bad new player experience, which then increases the problems all players suffer when a player base shrinks over time.

If the advantage of a veteran player in a multiplayer PvP game over new player was only skill, that would still be acceptable. But as most games have other forms of progress, an aging PvP multiplayer game ends up with veteran players beating up new players in matches that wouldn't even be fair at equal skill. Even I, and I am really not a very good player, can get excellent results with low tier tanks, just because I put equipment like binoculars and camo nets on them, and a 100% skill crew. I mostly play through low tiers quickly in order to reach the higher tiers of a tank line, but if I wanted to pad my win stats I could just farm noobs to the detriment of the overall game.

I was wondering whether PvP multiplayer games shouldn't have reverse progress: The more you play, the less strong your characters / vehicles / whatever would become. A veteran would have the increased skills of playing a lot, but not an already more powerful starting position. Instead he would have weaker stats than the new players, which would make the game a lot more fair. Playing a lot gives you medals and status, but would decrease your power. When you think you have become too weak, you could always start over, losing both the status and the handicap.

Labels:


Monday, April 29, 2019
 
The weak hitting the strong

To some extent the combat system of World of Tanks resembles the combat system of Dungeons & Dragons: A first random roll determines whether you get past the armor, then a second random roll determines the damage you do. However where the two systems differ a lot, is the power curve: How effective are weaker participants against stronger ones. In D&D 5th edition the power curve is relatively flat; even a weaker monster still has some chance to hit and damage a higher level hero. In World of Tanks the power curve is very steep; you end up facing tanks you can't possible damage, at least not while they are in a typical defensive position facing you.

The steep power curve has an especially strong effect on heavy tanks. If your enemies are two tiers lower than you, you could be an invincible god. Next match, in the same tank, against enemies two tiers higher than you, your armor might as well be made out of cardboard, and your gun becomes ineffective. There is no way to have a tank on which the armor is still effective against enemies two tiers higher. But you can play tank destroyers whose guns will still work reasonably well against higher tier opponents. That makes your performance a bit more consistent, and you suffer less from the inevitable bad match-ups.

One of the probably unintended consequences of the steep power curve is the effect on premium ammo. Some tanks have guns that are good enough with their regular AP or HE ammo to take on any comers; other tanks simply become worthless against higher tier opponents if they don't load premium ammo.

I must say that from a game design point of view, I much prefer the flatter power curve. Of course a weaker opponent should have less chance to hit, or deal less damage. But if you consider that for matchmaking you need to be the weaker opponent sometimes, the game remains more fun if you can at least still do something instead of ineffectively bouncing against too heavy armor. Right now I'm playing a lot of tank destroyers because of that.

Labels:


Sunday, April 28, 2019
 
Out of the Abyss - Preparing chapter 14: The Labyrinth

As told in my ongoing journal of my Rage of Demons campaign, my group just finished chapter 13 of Out of the Abyss, and I am preparing chapter 14 for the next session. Unlike the previous chapter, which I mostly played as written in the book, I find that chapter 14 has multiple serious problems, and I will have to rewrite a lot of it.

The most pervasive problem of chapter 14 is that it doesn't appear to fit with the level that the characters have at this point. My group is level 11, and as the book says that they should have been level 10 in chapter 11, level 11 seems perfectly working as intended. However most of the random encounters as well as the scripted encounters involve monsters like gnolls (CR 1/2), minotaurs (CR 3), cultists (CR 1/8), ghouls (CR 1), gargoyles (CR 2), and the like. And not in numbers that would even remotely challenge a level 11 group, even without the NPC army they have with them. While an occasional low challenge encounter can be fun to show the group how strong they are, a series of them would be boring and tedious.

The biggest event in this chapter is not much fun as written either, at it is very obviously scripted and doesn't involve any player agency. They come upon a demon lord CR 24 fighting another demon CR 17, and are supposed to just wait until the lesser demon is slain and the demon lord left, so they can recover the heart of the demon. If you play that as written, it becomes just scripted narrative without the players having any input or feeling any pride in the outcome. Combined with the low level of the other encounters in the chapter, the players might well misjudge the power of the demon lord and attack him. Any number of things could go wrong with that if you just spring that encounter on the players as a surprise. I think it would be a lot better if the group is aware of both hunting parties, the one led by the demon lord and the one led by the demon. They can then decide what they want to do about them, attack them separately with good preparation, or get involved in the two groups meeting and fighting each other. If the group themselves stages the fight between the demons, they would feel much better about the outcome.

The last event of the chapter is such a problem that I decided to completely cut it out. It involves a device called the Maze Engine, which is designed to produce a series of random effects, which are quite serious. Players could be thrown back in time to the start of the adventure, having to play through everything again. Or they could lose all magic items. A character might get disintegrated. While there are also some beneficial effects, overall the risk is far too high that if played as written a random result completely throws the campaign out of whack. Again this is a problem of player agency: They can't know the possible effects of the engine, and they can't easily shut the engine off once activated. And other than messing with the group, the event doesn't have any real function within the overall story.

So now I am rewriting much of this chapter in order to make it both interesting and playable. But I think the underlying problem is that D&D in general gets weird in the second half of the level curve. "Random encounters" that challenge a level 11 group don't make much sense, as the monsters you'd need for that were designed more like boss mobs. With a group that powerful, it is better to just skip the small stuff, and only really play through the main encounters of the book to the end.

Labels:


Saturday, April 27, 2019
 
Playing for the game vs. playing for progress

I am currently having much fun in World of Tanks, working on the two German tank destroyer lines, still at low tier. One of the fundamental problems of WoT is that it gets worse when you progress. There is no separation of players by skill at all, any skill level can play battles at any tier in random battles. While there are experienced players "seal clubbing" in low tiers, the highest concentration of experienced players is in the high tiers. So while Wargaming will happily sell you tier 8+ premium tanks or make your progress into the high tiers easier with Top of the Tree events, actually playing in tier 10 games against people who have played for years and are so much better than you isn't much fun. I'm not even playing my recently acquired French heavy tier 10 tank anymore, although I would still get extra experience and bonuses for doing so until the end of the month.

While World of Tanks does have "raid-like" content for more experienced players, that content is only accessible in groups and not sufficiently attractive to keep the veterans out of random battles. For both official xp/credits and unofficial WN8 rating, a point of damage is a point of damage, regardless of whether the tank on the receiving end was a noob playing a tank two tiers below you, or whether there was actually a challenge involved in the fight. Thus if you are an experienced player, your best bet is to play your tier 10 tank in random battles, as you always will be top tier. Less experienced players with tier 10 tanks will still be top tier, but won't actually do so well against the veterans. So for me the best strategy is to play tank lines only until tier 7, which even with +2 matchmaking will never see tier 10 tanks.

Of course that means that in a game with a significant part about progress, gaining xp and tiers in tech trees, I don't even want to progress beyond a certain point. Which makes that progress less interesting than in other games. Even if you didn't want to raid, you still would want to reach the level cap in World of Warcraft. Not so in World of Tanks. For the players that has the positive side that progressing to tier 7 is relatively fast, and the much slower progress to tier 10 is better skipped anyway. However for Wargaming that concept isn't so attractive, because what they are actually selling to paying customers is faster progress. If too many average players would realize that the top tiers are not worth getting to, they would stop paying for faster progress there.

One weird side effect is that I care very little about the progress I made before, my tier 8 to 10 tanks. If for some reason my account was hacked or a server bug would erase all my progress, I wouldn't even shed a tear. I could have pretty much the same fun with a brand new account.

From a game design point of view, that isn't ideal. If I had a say in design, I would do two things: Make rewards dependent on the tier of tank you damaged, possibly even on the rank of the player driving that tank, to make bottom feeding less attractive. And create some more attractive "end game content", like random battles for people with a certain personal rating or number of battles played, to keep the players with years of experience separate from the new and average players. That would give the experienced players more of a challenge, and open up more interesting progress for the average player.

Labels:


Thursday, April 25, 2019
 
Rage of Demons: Session 13

In the previous session the group learned of a way to banish all those demon lords back into the abyss. However this will require gathering a list of ingredients, which they now set out to do. First stop: The Wormwrithings, an area of the Underdark dug out by purple worms. Vizeran had told them of the location of a purple worm nest, from which they needed an egg [By pure coincidence we played this on Easter Sunday], and of a beholder's lair, from which they needed the beholder's central eye.

However on the way there they first encountered a booming voice in the dark, warning them to turn back. This turned out to be a minor illusion, cast by a lost dark elf young woman, who had gotten separated from her mother and her group of purple worm egg hunters. While the rest of the group wasn't much interested, Arkoy insisted of helping her to get back home. After some mishaps he managed to teleport her close to Menzoberranzan, gaining the gratitude of House Hallen there.

Getting closer to the nest, the group experienced earthquakes caused by burrowing worms. This led a part of the floor on which they were standing to collapse, landing them in some caves. This was in fact part of the worm nursery, and started the clock on a time table I had prepared. The first event on that table was a purple worm passing them, opening up a tunnel to the cave where the eggs were. The eggs were glued into a web, and had to be hacked out. While they were doing that, a group of dark elves turned up. Those were of course the mother of the dark elf they helped earlier, which led to a truce between the two groups.

While each group was getting eggs from a different corner of the web, another purple worm turned up. While the group could have avoided a fight by using the worm's blindness (which is what the dark elves did), they decided to fight it instead. They won the fight, but lost a part of their army. Before another purple worm could turn up, they grabbed two eggs and left. On they way out they encountered a fomorian giant, but that one turned out to be not much of a challenge.

So the group went on to the beholder's lair. There they first encountered some slaves, who weren't of much help. Then they met a mage who was acting as a majordomo for the beholder, holding a staff with an eye on top of it. They decided to kill him, which was quick enough, but Surina got hit from a death ray from the eye staff, and some more troops got killed in a cone of cold. Finally they reached the beholder's central lair, a big shaft criss-crossed by bridges. The floating beholder was not easy to reach. When Arkoy and Mog'burz teleported onto a bridge right next to him, the beholder used his disintegration ray to destroy the bridge, so they fell 500 feet and suffered 20d6 of falling damage.

The beholder also used his anti-magic cone to good effect, blocking magic from one direction while firing his eye rays into another. However he couldn't block everything. Quarion the ranger dealt a lot of damage with arrows, and some spells also went through from the group casters while the beholder was looking in another direction. The fight was truly epic, although I forgot to use the beholder's legendary lair actions. Finally they managed to kill the beholder, and we ended the adventure there, with the group having reached level 11. The Wormwrithings is a really good chapter in this adventure. The next chapter, The Labyrinth, will need some work to make it as good.

Labels:


Friday, April 19, 2019
 
Top of the French Heavy Tree

I am happy to report that I reached the AMX 50 B, the tier X tank of the French heavy line in World of Tanks. I'm way early, the Top of the Tree event goes until the end of the month, but I had bought a €10 boost that gave me the 5x for victory XP not just once per day, but three times per day. That does speed up things a lot, and so only tier IX of this event felt a bit like a grind.

However I still have mixed feelings towards the French heavy tanks with their autoloader guns and paper-thin armor. Sometimes one feels a bit persecuted when driving one, as enemy tanks will ignore other tanks and shoot you first whenever they see you. Not just because every shot on a French heavy penetrates due to the armor being negligible for this tier, but also because of the threat it poses. In one of my tier IX games the enemies had all decided to rather shoot at my team mates than me. In that game I did some insane damage, twice the hit points of my own tank. If you can unload a full clip into the side or back of an enemy tank, it gets devastating pretty fast. So I gained some understanding why everybody seemed to seek me out for destruction.

The result of all this is that while me battle results in World of Tanks are always somewhat inconsistent, this volatility goes way up when I play French heavy tanks. Some very good games, but a lot of games where I die before reloading the second magazine. On average the Top of the Tree event lowered my overall WN8 rate, because I had more very bad games than good games. I know that the trick is to stay behind and act as support, but in a random battle that doesn't work so well, because you can't rely on the front tanks doing a good job "tanking" for you.

Labels:


Tuesday, April 16, 2019
 
World of Tanks Challenge Events

The "American Challenge" event finished yesterday on the European World of Tanks servers. The basic premise is that you can get a premium tank for free if you grind a large amount of missions during 10 days. The grind has 10 stages, and each stage is harder or grindier than the previous one. Quickybaby, a World of Tanks YouTuber/Twitch streamer and top level player tried the previous challenge on his free to play account, and it took him 60 hours. Not a casual-player friendly event, or is it?

As it turned out, I kind of misjudged this. I completely ignored the event, and concentrated on the Top of the Tree event instead. But then I noticed that I was getting stage rewards from the American Challenge event from time to time. Because while the Challenge is a grind, it isn't an *extra* grind. No, if you just play for some other purpose, or even just for fun, as long as you are using at least a tier VI tank, you slowly accumulate progress towards the American Challenge.

At the end of the Challenge yesterday, without actually having done anything for it, I was at stage 7 out of 10. Which means I got 7 packs of stage rewards without any extra effort. And then I realized that the event was even better than I had thought: While you need to finish all 10 stages to get the premium tank for free, each stage counts for a 10% discount. So I ended up buying that TS-5 premium tank for 70% off, which made it really quite cheap. And its a really good tank, a bit slow and cumbersome, but with very good frontal armor and gun, which makes it an excellent face-to-face brawler.

Which brings me to a sidenote regarding armor in World of Tanks: It works better than the math would suggest. I realized that while playing through the French heavies Top of the Tree event, because I found that these lightly armored French heavies are worse than the math would suggest. Whenever an enemy had the choice between shooting me or shooting the tank next to me, that enemy would choose to shoot me. That is a consequence of the reward system of World of Tanks, both the official one and the inofficial WN8 rating system: They both reward damage dealt above anything else. If you fire a shot at something heavily armored and don't penetrate, you just took a big loss to your credits earned, xp earned, and WN8 rating. So the logical consequence is that everybody first tries to shoot anything that has light armor before aiming at anything well armored. In terms of game design that is bad, I would say the reward structure of World of Tanks is pretty bad. But if you have a heavily armored tank, it is an advantage.

I will still try to finish the Top of the Tree event this month, as I am already at tier IX. But overall it turned out that the American Challenge was actually the event that gave me a tank that I will continue playing. The French tier X heavy is probably not going to be played much, and I already sold the tier I to VIII ones, and will sell the tier IX once I got to X.

Labels:


‹Older

  Powered by Blogger   Free Page Rank Tool