THE NOT-SO-GREAT GAME OF THRONES: ASCENT ZOMBIE APOCALYPSE ANTICLIMAX HOWARD A. LANDMAN HOWARDL11 1. The Game Game Of Thrones: Ascent is a browser Flash game based on the popular HBO fantasy series. The game has many incomprehensible design features and numerous bugs, many of which have remained unfixed for years. I started playing in 2014, but in early 2015 I began a systematic exploration of certain bugs and how they interact. This note will detail the bugs and how I exploited them. 2. You and your sworn swords Except for building and crafting, most actions in the game involve using your Sworn Swords. These are non-player characters that you control. Each SS starts at rank 1 with 0 training points; as they gain experience from adventures and other activities, they go up in rank and gain 1 training point per rank. Early ranks are relatively easy to gain but later ones are much harder; going from rank 1 to 2 requires 10 XP but reaching rank 127 (the maximum) requires a total of over 1.2 million XP. Rank and TP per se do nothing to make your SS stronger. You need to spend the training points on either base stats or a specialty. The base stats are Battle, Trade, and Intrigue; Each SS comes with a starting bonus to one of these, which can be 30, 14, 6, 2, 1, or in a few special cases 0. Each TP spent on a base stat raises it by 2. Every stat also has 3 specialties. Battle has Fight, Harass, and Aid; Trade has Barter, Swindle; Intrigue has Spy, Sabotage, and Steal. When your SS reaches level 3 or greater, you are not allowed to train base stats further until you have trained in a specialty; once you have chosen a specialty, you may not train any other specialty. Each TP spent on the specialty gives that SS a 1% boost to attack and defense of that type. Although in theory an SS could have more than one base stat raised, this is rarely done because all individual actions are in a single specialty. Your main character also has base stats and % boosts, and these are added to those of the SS. So, for example, strength in a Fight action is given by (main battle + ss battle) (100% + main fight% + ss fight%). When losing an action, your SS may receive 1 to 3 wounds. These may be healed with medicine or, if left alone, heal by themselves at the rate of 1 per hour. An SS that accumulates 5 wounds is dead. A dead SS cannot perform any actions, but may be replaced by paying silver. The new SS is the same as the old one, except that: Their name is changed. 1
2 HOWARD A. LANDMAN HOWARDL11 They lose one rank (unless they are rank 1). They lose one percent in their specialty or 2 points in a base stat (but will not go below 0). The new SS is not available for use for 24 hours. Starting bonus stat points are exempt from this and cannot be lost. The obvious conventional wisdom conclusions from the above are that higher rank is better, SS that start with more bonus stats are better, no SS can get more than 126 training points, you should spend training points as soon as you receive them, and having an SS die is bad. It took me a very long time to realize that all of these were completely false. 3. The Training Bug The Training Bug is a bad name, because it does not actually occur while training, but that is what it became known as in the GOTA community. In fact, I already presented it in the previous section, where I described what happens when a dead SS is replaced. Did you see it? The bug is a simple programmer oversight. They did not consider the possibility that an SS would have gained training points but that none of them would have been spent on stats or %. In that case, the SS still loses one rank, but no stats or % are removed because they are all 0. Thus an SS can keep all its TP and go down a rank, at the cost of dying, paying silver, being replaced, and waiting a day. Considering rank and TP as separate variables, we see that higher rank is actually harmful. The only effect of increasing rank is to make TP harder (and eventually impossible) to earn. Thus, holding TP constant, lower rank is always better and being rank 1 is best. It should also be obvious that the bug removes the limit of 126 max TP. The benefit can only be realized if no TP have been spent, so one should never spend any TP until an SS is as strong as desired. And finally, death is devoutly to be wished, as it is the only path to ultra-strong SS. Despite the fact that this bug could be fixed by adding one line of code (checking if there are unspent TP and removing one), and that I explained this to the developers, the bug has persisted for over 2 years and is still active. 4. Early Exploits Given only the training bug, it is not possible to rapidly raise the TP of a single SS, due to the one-day cool down. There are items that reduce this delay slightly, and in fact one such item was given out free in a World Event just when I realized how useful they would be. With the 6% reduction from that, replacing took a little over 22.5 hours. I began to save gold to buy similar items, but they were expensive and not always available in the store. Since the cool down only applied to one SS, though, multiple SS could be replaced in parallel. The only limit to this would be how many SS could you kill in a day. I started garrisoning about 20 SS in Alliance-vs-Alliance (AvA) camps; if a large enemy attack came
THE NOT-SO-GREAT GAME OF THRONES: ASCENT ZOMBIE APOCALYPSE ANTICLIMAX 3 in, several of them could die. However, this was unreliable, and I never got more than 11 deaths in a single day. When AvA wasn t running, this dropped to zero. I clearly needed a more effective way to kill my SS. 5. Bread For Death One option was to send SS on adventures. Unfortunately, another bug got in my way: because the space on an SS s card is limited, if the SS was not fully trained a Train button would appear and push the Adventure button off the page. Thus it was impossible to send an untrained SS on a normal adventure. Fortunately, it was still possible to recruit such SS for an Adventure Party (AP), where multiple SS get sent out together. If the SS was bad enough (for example, a trade SS with no battle sent on an adventure using Fight), there was a good chance that it would lose and get at a wound. The problem is that those wounds could not just be saved up; wounds heal at a rate of 1 per hour. So, while I was struggling to kill an SS, the previously-acquired wounds were slowly evaporating. This meant that I might have to wound an SS 6 to 10 times before it reached 5 wounds. An improvement to this nightmare is to spend bread to speed up the adventure. Each bread cuts 30 minutes off the adventure time. However, bread take 1 hour to make, and is useful for speeding up other things like quests as well. On average, I found that it took 10 to 20 bread to kill an SS quickly. This was not something that I could sustain. 6. Instant Completion I get a lucky break when another player told me about a timer bug that allowed instant completion of adventures. It works like this: First, you get your SS to join another player s Boss Quest (BQ). (Alliance Challenges (ACs) also work.) BQ opportunities arise frequently, but because so many players want to join them in order to level faster, most of them are over within 30 seconds of being posted. Considering that it takes me 18 seconds to join one (which requires clicking on a URL and hence starting an entirely new instance of the game!), many people who try to join one find it already over before they can get in. However, assuming you succeed, then you do not want to see it complete. Instead, close the BQ window and wait until other players complete it. At that point, your SS will become free again, but you will not have collected any of the rewards. Join your AP with that SS and send it on an adventure. As soon as it is engaged, reopen the BQ to collect your rewards. The code for BQ finish has a bug that it deletes any adventure timer running on your SS as well; thus, the adventure completes immediately and any rewards and wounds can be collected. This was a great improvement over the bread method, but getting into BQs was chancy; on some days I was only succeeding one time in ten. And although getting into ACs is easy and reliable, they can take hours to complete, and so tend not to provide a practical speedup. Still, with decent luck, I could get an SS to have a complete-but-unviewed BQ every 10 minutes, and therefore kill it in under an hour without using any bread.
4 HOWARD A. LANDMAN HOWARDL11 7. The Lord Roxton Wound Factory I thought long about the best way to wound SS, which now appeared as one of my two primary bottlenecks. (The other was the 24 hour replacement time.) I eventually realized that the first hard Lord Roxton quest, with difficulty 100 and 10 minute duration, was ideal for my purposes. For people starting the game, the Roxton quest is usually the first really hard one, one which you must beat to progress, because if you lose it you just get it over again. Most people hate Roxton because of this, and celebrate when they finally win. But I was about to embrace Roxton and discover yet another helpful bug. Unfortunately, I was already far too strong and had already beaten Roxton. There was no way to go back but to reincarnate. After reincarnating, I decided to lose to Roxton using Fight, which meant I needed to keep my Battle at close to zero. (In theory there should have been no need to keep my Fight percentage low, because any percentage times zero is still zero, but most things that add Fight % also add base Battle stat.) This was incredibly painful, as it meant I could not upgrade any of my Battle-related buildings, or have any equipment on my SS or main character that gave even a few Battle points. This prevented many types of crafting and also severely inhibited leveling. However, submitting to that discipline, I finally reached the Roxton quest with 0 Battle. Although it had a nominal duration of 10 minutes, I had gear (and some level-30 Fight speedup talents) that reduced that by 48% to 5 minutes 12 seconds. Combining that with the free 5 minute speedup meant that I count finish a Roxton quest in 12 seconds plus however long it took me to click all the buttons. In practice I found that I could finish 5 Roxton quests and replace the now-dead SS in under 2.5 minutes (ignoring the 24 hour timer). However, this would not work for SS that had any Battle strength, because they might actually win. The game was designed with the intent that, no matter how strong you are, you always have at least a 5% chance of losing, and no matter how weak you are, you always have at least a 5% chance of winning. This would be painful for using Roxton to kill SS, since it would mean that on average you would only get about 19 wounds before beating Roxton (which would require another reincarnation to reset him). Fortunately however, the implementation of this policy is buggy. When attacking Roxton with strength 0 or 1, the game says that you will win on rolls of 1 to 5 and lose on rolls of 1 to 100. Clearly these can t both be right; in fact, you lose 100% of the time. It is still possible to beat Roxton by accidentally clicking on the wrong attack type, but this happens rarely. In my longest unbroken run, I lost to Roxton over 2700 times in a row. 8. Instant Karma At this point, I was able to reliably kill an SS in 2.5 minutes, but then still had to wait close to 24 hours for it to be replaced. So although I could kill many SS per day, each SS could only die once per day. However, I noticed that the cool down period was implemented with a timer that looked identical to a quest timer. I wondered whether the
THE NOT-SO-GREAT GAME OF THRONES: ASCENT ZOMBIE APOCALYPSE ANTICLIMAX 5 waiting period might be implemented as a kind of quest, or at least share the same timer code. If so, then the Instant Completion bug might apply to it. It only took one experiment to show that this was in fact true. Viewing a completed BQ would reset the replacement timer on a dead-but-reviving SS, saving nearly a day of wait time. Making use of this Instant Karma variant of the Instant Completion bug, I was now able to kill he same SS many times in a day, following the procedure: Have the SS join another players BQ, but close the BQ window before it finishes. Send the SS on the Roxton quest with 0 attack strength. Use the 5 minute free speedup. Get a wound or two. Repeat until dead. Pay to replace the dead SS, starting the 24-hour replace timeout. Wait until the BQ has finished (if it hasn t already). View the finished BQ, which clears the timeout and gives you a usable SS again. The hardest and most uncertain part of this is joining the BQ, which may need multiple attempts and take 5 or 10 minutes or longer. All the other steps can be completed in under 3 minutes. 9. Grinding It would be almost as boring to relate the subsequent grind, comprising over 4000 wounds and 800 SS deaths, as it was to perform it. So I won t. If I had to do that over again, I d probably want to automate it. Eventually, I had one SS with 555 training points (which proved that the TP variable was at least 10 bits long), as well as two others with more than 126. So far, everything I had done was relatively innocent, merely a set of science experiments. I could have continued in that vein, with the zombie SS untrained, and just continued to accumulate more TP. But I was also curious about how overpowered SS would perform in the game; for that, I had to train at least one of them up in real stats and a specialty. 10. Training, Testing and Tales I decided to train the highest SS in Trade/Swindle.