Riddler - Solutions to Castles Puzzle: castle-solutions-3.csv
Data license: CC Attribution 4.0 License · Data source: fivethirtyeight/data on GitHub · About: simonw/fivethirtyeight-datasette
264 rows where Castle 2 = 0 sorted by rowid descending
This data as json, copyable, CSV (advanced)
Suggested facets: Castle 1, Castle 3, Castle 4, Castle 5
Link | rowid ▲ | Castle 1 | Castle 2 | Castle 3 | Castle 4 | Castle 5 | Castle 6 | Castle 7 | Castle 8 | Castle 9 | Castle 10 | Why did you choose your troop deployment? |
---|---|---|---|---|---|---|---|---|---|---|---|---|
1321 | 1321 | 1 | 0 | 0 | 0 | 0 | 0 | 10 | 27 | 29 | 33 | My focus was on getting 28 total victory points out of a possible 55, so I concentrated on 8, 9, 10, and winning 1 extra point on the "1" castle. |
1320 | 1320 | 0 | 0 | 3 | 3 | 16 | 6 | 16 | 21 | 4 | 31 | I know this is really late, but here is a serious entry. The code used to generate this is at https://pastebin.com/ieFeGQzN |
1318 | 1318 | 0 | 0 | 0 | 18 | 18 | 2 | 2 | 2 | 34 | 24 | This strategy beat the previous top-5. |
1295 | 1295 | 1 | 0 | 0 | 26 | 1 | 1 | 26 | 26 | 17 | 2 | The simplest win is on 10/9/8/1. Two problems: it's already popular, and weak players over-defend Castle 10. I'll try to win on 9, 8, 7, and 4 instead. |
1291 | 1291 | 2 | 0 | 5 | 10 | 0 | 0 | 24 | 24 | 35 | 0 | Trying to focus on getting 28 victory points while sacrificing the "10" assuming most people will want the big win. |
1269 | 1269 | 0 | 0 | 11 | 14 | 0 | 21 | 25 | 29 | 0 | 0 | I’ve narrowed down the gameplay to around 14 possibly optimal plays. This is one of them. There are 33 possible exactly 28points to win strategies. This one is 8-7-6-4-3. Allocated by relative castle value. Castle/28*100. Here’s the list of 9, allocate by taking castle/28*100: 10-9-6-3 10-9-5-4 10-8-7-3 10-8-6-4 10-7-6-5 9-8-7-4 9-8-6-5 10-6-5-4-3 8-7-6-4-3 The other 5 are semi suboptimal vs the 9 but forms the “rock,paper,scissor”: ExpectedValue: castle/55*100 EvenAcross: 10/castle Ultimate: castle/28*100+1 for castle 8,9,10 Lucky7: castle/28*100 for castles 1 to 7 Troll: 47,53 on castle 9 and 10 respectively. At least one of these strategies will do well depending on the market. And the market will shift around these strategies depending on the amount of trolldom. |
1266 | 1266 | 4 | 0 | 5 | 0 | 2 | 2 | 18 | 30 | 20 | 19 | Developed a troop deployment that beat 1386 out of 1387 of the castle-solutions.csv from two years ago. |
1259 | 1259 | 0 | 0 | 0 | 0 | 0 | 0 | 12 | 12 | 12 | 64 | focused highly on the highest valued castles |
1251 | 1251 | 0 | 0 | 0 | 0 | 0 | 14 | 14 | 14 | 33 | 25 | Looking at previous results the middle became the highest value for least deployments. BUt I wanted to be able to take the the 10 and 9 as well. so I loaded the top end and placed enough in the middle that might get me to 28 points. I am willing to cede 15 points to the opponent |
1250 | 1250 | 0 | 0 | 13 | 13 | 14 | 14 | 12 | 12 | 11 | 11 | I'm figuring that most people will concentrate there forces mostly in the first few castles and somewhat in the last few. With this strategy I think i'll have a strong troop advantage in the middle castles and a weaker troop advantage in the end while only completely ceding the first 2 castles. Even if someone uses a similar strategy with a single troop in the first 2 castles, I'll still have a competitive advantage in at least one castle without sacrificing a more dominant position in the middle and end. |
1241 | 1241 | 0 | 0 | 3 | 5 | 11 | 13 | 21 | 22 | 14 | 11 | Kind of a guess, really |
1240 | 1240 | 5 | 0 | 0 | 12 | 0 | 13 | 0 | 30 | 35 | 5 | Trying to secure a baseline of 17 and steal either 10 or 7+3 as well as the first castle |
1238 | 1238 | 3 | 0 | 6 | 8 | 15 | 22 | 4 | 3 | 31 | 8 | a computer told me to |
1237 | 1237 | 0 | 0 | 0 | 10 | 10 | 25 | 25 | 15 | 10 | 5 | Because the middle will be ignored |
1233 | 1233 | 0 | 0 | 0 | 16 | 21 | 1 | 2 | 1 | 35 | 24 | Optimised against top fives from both runs and median from the first. Depends on snatching the top two bolstered by four and five, these four wins would total a bare minimum of 28 of 55 points. Sometimes snatches the 6–8. If most strengthened the top prizes a bit, yeah, I'm screwed. Didn't want to do a deep dive into the complete data. |
1223 | 1223 | 3 | 0 | 0 | 0 | 0 | 0 | 0 | 31 | 33 | 33 | all or nothing |
1222 | 1222 | 0 | 0 | 0 | 1 | 17 | 23 | 28 | 3 | 4 | 24 | Castle 8 and 9 are highly contested, so you have to put in a lot of troops to gain a high probability of winning them. However, if your strategy is 9-10 heavy, 8 is weak for you and I might win or tie with a few there; if your strategy is more focused on 8-10 or lower values, I might snag a tie or win with a couple troops in 9. Overall, the winning strategy is 5-6-7-10. If I lose 10, I hope to win 8 or 9, and tie or win a few of the lower ones. I will definitely lose games, but the hope is that I can win against a bunch of strategies. For instance, this beats about half of last years' winners. |
1218 | 1218 | 0 | 0 | 10 | 0 | 0 | 0 | 10 | 20 | 25 | 35 | The focus is on on reducing the battlefield down to enough castles to get 28 victory points, and then identifying the set of castles that make up 28 points that past players have shown the least interest in competing for. |
1213 | 1213 | 0 | 0 | 0 | 2 | 20 | 18 | 2 | 24 | 32 | 2 | Since the previous contest winners all focused on a group of castles totalling 28 points, I somewhat randomly chose 5, 6, 8, 9 and put 3 troops per point value in each of these. That left me 16 troops. I decided to minimally defend castle 4, 7, and 10 with two troops each and then reinforced two of my targeted castles with five more troops each. |
1205 | 1205 | 0 | 0 | 0 | 0 | 0 | 0 | 10 | 20 | 30 | 40 | Higher value=more soldiers, keep it simple |
1204 | 1204 | 0 | 0 | 11 | 11 | 1 | 20 | 22 | 34 | 1 | 0 | Trying to win the lowest number of castles that reach 28 points, with maximum force at higher numbered castles where more enemy attacks can be expected. We hope to take away castle 8 from anyone who is focusing on the top castles, and win some cheaply. |
1200 | 1200 | 0 | 0 | 0 | 5 | 12 | 16 | 18 | 24 | 25 | 0 | The top one and bottom 3 are simply not worth the manpower. |
1196 | 1196 | 10 | 0 | 0 | 0 | 0 | 0 | 0 | 25 | 37 | 28 | To win just over 50% of the points with the least number of castles by deploying enough troops to four castles to win 28/55 points and abandoning the other six |
1195 | 1195 | 0 | 0 | 1 | 1 | 22 | 24 | 1 | 4 | 25 | 22 | I wrote a half-baked genetic algorithm that evaluated strategies against random strategies, entries from the previous contests, and the top strategies from the previous generation, and then chose the strategy that most often received the highest fitness of its generation. |
1187 | 1187 | 0 | 0 | 3 | 1 | 12 | 17 | 2 | 31 | 2 | 32 | Well, the first time the winners targeted 7 and 8, and the second time the winners targeted 9 and 10. So I'm going to target 8 and 10 - as long as I win those and break even on 1 through 6, I should beat the copy cats from last time, and anyone who hopes to beat the copycats by one-upping them on key castles. In order to break even or better on 1 through 6, I'm targeting 5 and 6. After that, I've got 8 armies left to split among the remaining castles, in case I lose some of the others. I ignore 1 and 2, which aren't worth much, in favor of taking advantage of those who leave some higher-value castles empty or close to empty. I also made sure that my solution beats most typical solutions (i.e. even splits, or assigning armies proportional to value), as well as most of the winner's solutions (although admittedly Jim Skloda's submission from the first time counters mine pretty perfectly). I also think it's worth going for numbers that are 1 or 2 mod 5, since many people will submit nice round numbers, as proven by the winning submissions from the previous contests. |
1184 | 1184 | 0 | 0 | 0 | 10 | 10 | 12 | 14 | 16 | 18 | 20 | I am anticipating others wasting troops on the low value targets, which I will abandon. I assigned troops to each other site based on their value alone, anticipating the others at this point would overthink and leave the high value targets undefended(but in an unpredictable way) |
1182 | 1182 | 0 | 0 | 7 | 3 | 5 | 2 | 16 | 17 | 28 | 22 | hope |
1179 | 1179 | 0 | 0 | 0 | 12 | 0 | 0 | 18 | 30 | 40 | 0 | 28 is the minimum number of points to win. I sent the least number to castle 4 because I anticipated that it would not need to be taken with higher numbers in most scenarios. |
1176 | 1176 | 100 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | I'm a warlord, yes, but all I really care about is myself. . . and I want a castle! If anyone stands in my way they will be sorry. |
1165 | 1165 | 0 | 0 | 0 | 15 | 20 | 20 | 20 | 25 | 0 | 0 | Figuring the enemy would over commit to the larger value castles. |
1163 | 1163 | 5 | 0 | 1 | 10 | 9 | 12 | 5 | 0 | 18 | 40 | I ran a program that simulated a thousand rounds of battles with 20,000 participants and made random updates to each strategy after each round based on how well the players performed on the previous round. This was the winner of the last round. |
1160 | 1160 | 10 | 0 | 0 | 0 | 0 | 0 | 0 | 30 | 30 | 30 | Adds up to 28 |
1157 | 1157 | 0 | 0 | 0 | 0 | 0 | 0 | 20 | 30 | 50 | 0 | Random Hunch |
1147 | 1147 | 0 | 0 | 0 | 10 | 15 | 16 | 17 | 18 | 24 | 0 | Best balance of middle points |
1145 | 1145 | 0 | 0 | 0 | 17 | 19 | 20 | 21 | 23 | 0 | 0 | Capture the middle |
1143 | 1143 | 0 | 0 | 17 | 1 | 1 | 16 | 3 | 3 | 29 | 30 | Last time, the top players fully settled on a 10-9-5-4 strategy. I think this time around, players are going to actually contest the 10 and 9 castles correctly - the fact that 7 and 8 have had more troops than 10 and 9 is just wrong & the meta has to trend away from that - but it may not quite get there. This 10-9-6-3 goes for the same number of points, and should beat a fair number of other approaches (e.g., 9874) while being pretty good against the classic 10-9-5-4 just because 29/30 is on the high end for 10 and 9. |
1141 | 1141 | 0 | 0 | 8 | 0 | 13 | 4 | 6 | 5 | 36 | 28 | Similar to last time's champion, optimised against first and second submissions and solutions optimised against them with more weighting given to the latter. |
1139 | 1139 | 0 | 0 | 4 | 17 | 21 | 2 | 4 | 5 | 32 | 15 | evolutionary ai found a better solution |
1137 | 1137 | 0 | 0 | 0 | 14 | 21 | 1 | 0 | 1 | 33 | 30 | This combo won 100 simulation rounds in a row using randomized, previous champs, and tweaks of previous round winners. |
1133 | 1133 | 0 | 0 | 2 | 5 | 17 | 5 | 17 | 17 | 33 | 4 | I want to win a number of castles. I tried to adjust for the adjustments people would make when comparing the two previous winners. |
1130 | 1130 | 0 | 0 | 0 | 0 | 0 | 0 | 20 | 23 | 27 | 30 | Determine the maximum number of castles that can be abandoned while still achieving net victory assuming individual victories at the remaining castles. sum(i, i = 1 .. 10) = 55, sum(i, i = 7 .. 10) = 34, sum(i, i = 1 .. 6) = 21. 34-21 = 13, therefore only castles 7-10 need to be won. Soldiers were distributed approximately proportionally to the point value of the castle, but preferentially rounding down for lower value castles and up for higher values. |
1122 | 1122 | 0 | 0 | 11 | 12 | 14 | 4 | 24 | 4 | 27 | 4 | I'm sending 4 to castles 6, 8, and 10 to try to beat anyone who sends only a few there. I then focus on winning castles 3, 4, 5, 7, and 9 to get my 28 points. |
1117 | 1117 | 0 | 0 | 0 | 0 | 0 | 17 | 18 | 18 | 29 | 18 | 55 points available. Give up the first 15 points and focus all the efforts on gaining by going above the average for each of the remaining castles. I went heavy on 9 assuming that most others would have the same thought process and skew towards the higher values except 10. |
1116 | 1116 | 0 | 0 | 1 | 6 | 11 | 18 | 28 | 5 | 0 | 31 | I tried to use Ken Nickerson's strategy from the first battle but with a focus on two castles that were differently successful in the first two battles. In the first one, 7&8 were the main targets by the top 5. In the next one, 9 and 10 became the big numbers to target. I need 28 points to win the battle. My goal is to take 5, 7, 6, and 10 in most matches. I get all four of those and I win. If I don't, well, hopefully I can steal the 8 (or the 4) and use dumb luck to conquer smarts. |
1107 | 1107 | 3 | 0 | 0 | 0 | 0 | 0 | 0 | 30 | 33 | 34 | I'm trying to get the majority of available points with the fewest castles. |
1098 | 1098 | 0 | 0 | 10 | 0 | 0 | 26 | 0 | 0 | 28 | 36 | No time = no thought = no analysis = no strategy. Anyone defeated by this should have a long walk accompanied by a bell and "Shame! Shame! Shame!" |
1097 | 1097 | 0 | 0 | 4 | 16 | 21 | 2 | 4 | 5 | 32 | 16 | variation on a theme |
1080 | 1080 | 4 | 0 | 9 | 5 | 1 | 18 | 1 | 33 | 3 | 26 | Attempted optimization against both of the previous two rounds. |
1071 | 1071 | 0 | 0 | 0 | 15 | 21 | 0 | 0 | 0 | 36 | 28 | Better than Mike |
1068 | 1068 | 0 | 0 | 0 | 15 | 18 | 1 | 1 | 1 | 32 | 32 | better than Derek |
1064 | 1064 | 0 | 0 | 0 | 15 | 15 | 0 | 0 | 0 | 35 | 35 | We go all in on the minimum value to win. |
1063 | 1063 | 0 | 0 | 0 | 0 | 8 | 2 | 25 | 30 | 35 | 0 | |
1057 | 1057 | 0 | 0 | 2 | 16 | 21 | 3 | 2 | 3 | 32 | 21 | variation on a theme |
1054 | 1054 | 0 | 0 | 0 | 5 | 4 | 5 | 24 | 5 | 30 | 27 | Winning the first few castles is essentially meaningless, so any significant troops sent there are wasted, even as a blocking action. Beyond that point, it's a matter of trying to strike a balance with remaining troops between attacking in force, and defending against small raids. There seems to be a consistent trend in the previous battles to focus most troops on Castle 8, so that seems to be the best place to not fight too hard over, in order to preserve sufficient troops to win other battles instead. |
1051 | 1051 | 0 | 0 | 3 | 0 | 0 | 14 | 14 | 5 | 33 | 31 | I tried to come up with a troop arrangement that would outscore the top five deployments (averaged out) and the top deployments from the previous rounds. It was mostly a matter of trial-and-error. And I didn't quite succeed in my goal (my deployment beats the "average" 36-19 and the second round winner 43.5-11.5, but loses to the first round winner 25-30). But I feel good about my choices of castles to attack with strength (9, 10) and about my decision to emphasize attacking castles 6 and 7 at the expense of castles 4 and 5. I am a little bit uneasy about my decision to make only a modest 5-troop deployment to castle 8 as there may be a rush by others to scoop up those points this round. But I think the decision to abandon castles 1 and 2 in favor of a token 3-troop deployment to castle 3 is sensible. |
1048 | 1048 | 0 | 0 | 5 | 18 | 20 | 1 | 25 | 26 | 3 | 2 | focus mainly on the the middle castes, sacraficing castles to increase distribution to castles 8,9 |
1045 | 1045 | 0 | 0 | 2 | 16 | 21 | 3 | 2 | 2 | 32 | 22 | Best of last two plus some ai |
1043 | 1043 | 0 | 0 | 0 | 5 | 5 | 15 | 15 | 15 | 20 | 25 | Random |
1042 | 1042 | 0 | 0 | 0 | 5 | 5 | 15 | 15 | 15 | 20 | 25 | Random |
1035 | 1035 | 0 | 0 | 0 | 3 | 0 | 22 | 23 | 24 | 25 | 3 | |
1031 | 1031 | 0 | 0 | 2 | 2 | 22 | 4 | 22 | 22 | 4 | 22 | |
1030 | 1030 | 0 | 0 | 1 | 15 | 1 | 1 | 26 | 26 | 30 | 0 | I just tried to ensure I had 28 points and didn't want to invest in 10 or 1/2 |
1027 | 1027 | 0 | 0 | 0 | 0 | 0 | 10 | 20 | 30 | 40 | 0 | Most people will try locking in 10, I'd rather let them spend their points since 9 is almost equal. Further it allows me to hit a few more relatively high value targets further down |
1019 | 1019 | 0 | 0 | 0 | 0 | 0 | 6 | 16 | 21 | 26 | 31 | |
1016 | 1016 | 0 | 0 | 8 | 0 | 3 | 0 | 31 | 9 | 9 | 40 | Noticing that in both prior rounds people have hammered the middle numbers or the top numbers, but not both, I wanted an allocation that would win outright at one of those values (31 on 7, 40 on 10) while also winning whichever of 8 or 9 opponents leave under-defended, and winning enough lower-hanging points to get to magic number 28. |
1015 | 1015 | 0 | 0 | 2 | 2 | 17 | 18 | 27 | 3 | 4 | 27 | Hold strong on 10+7+6+5. If I don't win one of these distribute enough to hopefully get lucky on one or two other castles. This strategy has better than 75% win percentage against previous rounds and beats 8 of the 10 top 5 competitors in the previous two battles. |
1014 | 1014 | 0 | 0 | 2 | 14 | 15 | 5 | 5 | 5 | 34 | 20 | I assumed that most people would choose a strategy from one of the top performers from the last time we ran this competition. I started my “strategy bank” with the top three performers from last time. Then, my process was to move a single soldier from one castle to another for each strategy, store this as a new strategy in the “strategy bank”, play each strategy against the others, and keep the top 2% performing strategies as the seed for the next generation of strategies. I coded this in Matlab. After 5 generations, the top strategy I got was [0 0 2 14 15 5 5 5 34 20]. |
1011 | 1011 | 0 | 0 | 0 | 15 | 2 | 3 | 21 | 25 | 30 | 4 | Figured this setup would get me the 28+ points I need against most other folks' deployments. |
1010 | 1010 | 0 | 0 | 0 | 0 | 0 | 0 | 10 | 20 | 30 | 40 | |
1006 | 1006 | 0 | 0 | 0 | 0 | 0 | 40 | 60 | 0 | 0 | 0 | Want to overwhelm the squishy undervalued middle with enough troops to fend off anyone who doesn't just flood one of the two castles. Pin the rest on luck and the fog of war. |
1004 | 1004 | 0 | 0 | 0 | 0 | 10 | 15 | 20 | 25 | 30 | 0 | sacrificed top and bottom |
1002 | 1002 | 0 | 0 | 0 | 18 | 18 | 8 | 5 | 5 | 35 | 11 | |
991 | 991 | 0 | 0 | 10 | 0 | 22 | 0 | 0 | 0 | 34 | 34 | I only need 28 points to win and castles 9&10 seemed undervalued by the average player. I’ve gone all in on four castles. |
982 | 982 | 0 | 0 | 2 | 3 | 12 | 15 | 18 | 11 | 5 | 34 | 28 to win. Win 10. Win any 3 of 5-8. |
981 | 981 | 0 | 0 | 0 | 0 | 0 | 25 | 25 | 25 | 25 | 0 | |
978 | 978 | 0 | 0 | 2 | 3 | 10 | 15 | 17 | 17 | 18 | 18 | Because in the last battle the most successful warlords targeted the middle and top numbered castles with an overwhelming number of troops, I wanted to spread my points more evenly across castles with a value of five or higher (because even if you conquer the lower castles you still lose). This general strategy might be susceptible to players who cluster their soldiers at the top, but I am hoping to split the difference and more evenly spread my troops in the hope that when the smoke clears I can - to paraphrase Varys from Game of Thrones - be king of the ashes. |
973 | 973 | 0 | 0 | 5 | 15 | 20 | 5 | 0 | 0 | 25 | 30 | God told me. |
967 | 967 | 0 | 0 | 0 | 6 | 12 | 18 | 26 | 32 | 3 | 3 | |
966 | 966 | 0 | 0 | 0 | 11 | 0 | 0 | 27 | 31 | 31 | 0 | No point putting a small number of soldiers in a castle as you get no points for a loss. 9+8+7+4=28 is just over half the maximum (55). I think a bunch of people will go all in on 10, 9, 8, 1 with a 30,30,30,10 spread and this will beat that. Similarly, this beats a 25-25-25-25 spread on 10,9,8,7 and the 10 on all castles approach. Finally by ignoring castle 10, we also beat the strategies that put alot on castle 10 and spread a little to everything else which I think might be common. |
964 | 964 | 0 | 0 | 8 | 11 | 15 | 18 | 22 | 0 | 26 | 0 | It looked about right. |
952 | 952 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 20 | 30 | 50 | Seemed smart |
951 | 951 | 0 | 0 | 4 | 5 | 17 | 16 | 25 | 0 | 0 | 33 | |
947 | 947 | 0 | 0 | 2 | 2 | 5 | 5 | 29 | 1 | 1 | 55 | winning #10 cancels out the first 4 if lost. then 567 > 89 so put more there. |
946 | 946 | 0 | 0 | 0 | 0 | 6 | 5 | 11 | 18 | 28 | 32 | Top Heavy |
945 | 945 | 0 | 0 | 2 | 3 | 20 | 20 | 20 | 2 | 2 | 31 | Castles 8 and 9 received a lot of attention in the previous two iterations, respectively, because of various assumptions about the other players. We’ll see if this will work, but 10/7/6/5 are enough to win, and I’m gambling on any deployment that beats one of those splitting other castles with me. |
938 | 938 | 0 | 0 | 0 | 5 | 10 | 10 | 15 | 30 | 20 | 10 | Just giving away the low point castles and loading up on the 8 and 9 but hoping to eke some wins out of the 10 and 7 |
927 | 927 | 0 | 0 | 12 | 1 | 1 | 23 | 3 | 3 | 33 | 24 | Used a genetic algorithm (the same as last competition) to explore distributions that would be good against the second round distributions and the first and second round distributions combined. Then used the same algorithm to optimize against *those* and the first and second round distributions simultaneously. |
926 | 926 | 0 | 0 | 0 | 0 | 10 | 12 | 15 | 18 | 21 | 24 | Started proportionally and then let go of the lesser castles |
925 | 925 | 0 | 0 | 0 | 20 | 20 | 20 | 20 | 20 | 0 | 0 | Why not? |
917 | 917 | 23 | 0 | 2 | 1 | 1 | 1 | 1 | 23 | 24 | 24 | I placed at least 1 troop to every castle except for 2. I assume that my enemy sends at least 1 troop to every castle and therefore will give me the best chance to win 3. Next I assume the point of the game is to get 28 as there a total of 55 points. By dividing up all other amounts amongst the quickest way to make 28, (10+9+8+1) I have given myself the best chance to win those numbers. |
910 | 910 | 0 | 0 | 0 | 4 | 4 | 10 | 17 | 28 | 32 | 5 | The additional deployment scheme was won with emphasis on castles 7 and 8 .. and in the reprise (second) simulation, the winning submission emphasized Castle #9 and #10. By putting 0 soldiers in Castle #1, 2 and 3, I am going to concentrate my forces in Castles #6 - #9 with just putting enough soldiers in Castle #10 to avoid giving it away cheaply. In addition, I am putting 4 soldiers each in Castles #4 and #5 as a way to score a few "cheap" points against people who concentrate almost exclusively in Castles #6 - 10. |
899 | 899 | 0 | 0 | 0 | 0 | 3 | 11 | 21 | 21 | 22 | 22 | Nothing complicated - just based on past winners and seems like an even mix across the top castles may work. |
898 | 898 | 0 | 0 | 1 | 2 | 21 | 14 | 3 | 33 | 4 | 22 | I ran a monte carlo with all the previous troop deployments, plus a bunch of variations on the previous successful strategy, and it popped out this trimodal distribution. Basically, I optimized a trimodal distribution to beat optimized bimodal deployments. |
896 | 896 | 0 | 0 | 0 | 0 | 19 | 23 | 0 | 27 | 31 | 0 | Go all-in on 4 castles that give just enough points to win (28), ceding the other 27 points’ worth. Stack a few more troops on the high value castles just because. |
894 | 894 | 0 | 0 | 0 | 0 | 18 | 22 | 22 | 33 | 0 | 5 | Give up 5 castles expecting to split points on some of them. Maybe get a cheeky 10 against similar strategies. |
893 | 893 | 0 | 0 | 1 | 1 | 15 | 20 | 20 | 1 | 1 | 41 | The most direct method of achieving a majority while (hopefully) limiting exposure to defeat by fielding more men along my prescribed victory path than does the opposition. No backup plan, no reserves. When in doubt, attack. |
890 | 890 | 10 | 0 | 0 | 0 | 0 | 0 | 0 | 15 | 25 | 50 | Forces concentrated on minimum four castles to win |
889 | 889 | 1 | 0 | 2 | 15 | 22 | 1 | 2 | 3 | 33 | 21 | |
875 | 875 | 0 | 0 | 0 | 7 | 23 | 5 | 4 | 3 | 34 | 24 | Beat the top player from last time then designed a strategy to beat that then designed a strategy to beat that |
873 | 873 | 7 | 0 | 0 | 0 | 0 | 0 | 0 | 26 | 31 | 36 | Protect the bag |
Advanced export
JSON shape: default, array, newline-delimited
CREATE TABLE "riddler-castles/castle-solutions-3" ( "Castle 1" INTEGER, "Castle 2" INTEGER, "Castle 3" INTEGER, "Castle 4" INTEGER, "Castle 5" INTEGER, "Castle 6" INTEGER, "Castle 7" INTEGER, "Castle 8" INTEGER, "Castle 9" INTEGER, "Castle 10" INTEGER, "Why did you choose your troop deployment?" TEXT );