A few thoughts. Manual rolling has to be true random. It simulates individual dice tolls. Which is true random. Unless they mess with how the dice are going to land. Also I use the slide bar when teaming up. So I roll 30 retreat then my ally rolls 30. Spike and Sterling actually have done quite a bit of the research and created calculators etc. Manual rolling can pretty dramatically change odds. I liked your video showing that website is great. I read a published article about a year ago and it was consistant with what you explained on the dice roll odds. Thanks for publishing this.
@@riskyfil8955 I noticed you use the bar to roll 3vs1 when you attack with your big stacks, why 3vs1 and not 4vs1? I think you need 4vs1 to be 100% sure of victory.
Here's a way to prove to yourself that the discrete dice rolls are not "Balanced". Rolling 1 die as attacker against defender's 2 dice is a guaranteed loss under "Balanced Blitz", but under "True Random" it wins 11% of the time. That means, if you EVER witness a 2-vs-2 where attacker rolls 1 die and wins, it is not balancing. Suggested test procedure: Start a game vs 2 bots. If not first, pause/restart until you are. You should get a couple 2-vs-2 shots per first turn. Just keep pausing/restarting until you win one. It will take 6 rolls on average. I've done it and won such a roll, proving discrete rolls are not balanced. The "?" help for the ambiguously labeled "Dice Rolls" game parameter clearly says it adjusts how "blitz" rolls are executed. The term "blitz" means resolving battle in a single action, as opposed to a sequence of user controlled rolls. They use the term "Balanced Blitz" to refer to their scheme that skews blitz results in favor of the likeliest outcome, reducing/eliminating the chances of long shot outliers, and in their view, better rewarding strategic play. The term "True Random" means resolving blitz rolls as an unseen sequence of dice rolls without any "balancing". When you choose to roll 3, 2, or 1 dice (as opposed to blitz), you are choosing to roll according to the classic rules of the Risk board game, regardless of the setting of the "Dice Rolls" game parameter. Bottom line, under "Balanced Blitz", the GMs that choose discrete dice rolls rather than blitz when chipping against a stronger defender are correct, gaining a small improvement in results. Interesting video. Thanks for posting.
The manual dice do seem to be true random at least from my limited experience. I usually use them to weaken an army before blitzing them. You have a lot more games played though so I'll have to check that.
Theres like no chance 0 that I'm as unlucky with the manual dice as what I've been getting I lose soo much attacking and rarely get ahead when starting at an equal stack small or big numbers and blitz I get destroyed almost every time while attacking it's not realistic at all and this is like 350 games
I won my first game (very first game) after watching one of your vids last night... And I found the Die/Dice Rolls to be very straightforward & there was VERY little NUANCE to the ' a 6 versus a 2 or 3 almost always won'. I just got 4th in this game this morning where EVERYONE except a starting A.I. (from a leaver) seemed to get very very very odd rolls in that consistently people were losing a 7 versus 2 or 8/9 versus 4 in the sense that '7-2 and the defender kept their territory (i.e. -6 troops for attacker and -1 for defender). It was happening again and again, it stopped 2 people from dying for about 3 turns even. I noticed in the second game that 'True Random Rolls of the Die' was on and so I haven't been sure if that's the cause of the difference. I do like True-Random though more I think in the end... The constant 'I know I'm going to crush an army completely and have enough to win by the time I get all the way across the map' is a bit much for me. I know each roll of a die in both modes can workout the same, so my losing 10 troops to an attacking army of 12 when they lost 1 troop is quite ridiculous even if not to the same degrees of ridiculousness that happens with your example of 20 losing to a 1.
Balanced Blitz affects a lot larger edge cases than you might realise. For example, the 4v1 has 100% on balanced, but its actual odds are only 91.15% (about 10 in 11). Considering how many players use a 4v1 to card trade most turns of the game, the certainty of success actually changes the flow of the game a lot.
I’ve about had it with the game. Feels like every time I’m attacking with a 20 against a 12 I somehow loose meanwhile also loosing while defending with a 100 against a 105… in balanced blitz.. makes zero sense. Great played games ending in the dumbest ways possible and consecutively, back to back.
@@1guard754 shame. I wish the game would allow you to make friends over steam so you can try to arrange a game with a bunch of players you know personally. Currently, you could only make risk friends which doesn’t facilitate communication.
Sounds like balanced blitz uses a pseudo random distribution. In essence it would be that first throw is a true random, and the second throw lessens the chance for that same outcome. In the 1 defending die case, first winning throw would have that 34,03 % chance, and second throw it would be less. How much less depends on the function..
No offense but I think you have manual roll wrong. I've used it multiple times to make a bigger army smaller than mine then blitz it and win. I've even used it against capitals which have 3 dice defence
@@MRSLIN Good point. It just seems like from experience that it's more random than skewed in the attackers favor because there's also been plenty of times where I lose manual roll multiple times in a row
@@Daimxn A way to truly see if it is still a balanced blitz or has changed to a true random is to conduct a test of hundreds or thousands of rolls with the manual, then compare it with the brute force data from data genetics. If it closely matches(meaning it is statistically significant), then it is true random; conclude otherwise if not.
This should not be a thing because it should be RNG. Except it is not.. I can prove it with a recorded replay. 6 was hit 1 time by defense in 28 rolls..
The dice kill me. They talk about how the numbers are even but I play hundreds of games every month and 80% of the attacks that I engage in the other person kills me.
A few thoughts. Manual rolling has to be true random. It simulates individual dice tolls. Which is true random. Unless they mess with how the dice are going to land. Also I use the slide bar when teaming up. So I roll 30 retreat then my ally rolls 30. Spike and Sterling actually have done quite a bit of the research and created calculators etc. Manual rolling can pretty dramatically change odds. I liked your video showing that website is great. I read a published article about a year ago and it was consistant with what you explained on the dice roll odds. Thanks for publishing this.
do you have a link for spike and sterling's analysis? google came up empty.
I'll ask Sterling.
They could easily change how the dice roll.
@@MarkkuS They do sometimes. It will get crazy sometimes while they adjust it.
@@riskyfil8955 I noticed you use the bar to roll 3vs1 when you attack with your big stacks, why 3vs1 and not 4vs1? I think you need 4vs1 to be 100% sure of victory.
I love you MR SLIN
Love you too haha
Here's a way to prove to yourself that the discrete dice rolls are not "Balanced". Rolling 1 die as attacker against defender's 2 dice is a guaranteed loss under "Balanced Blitz", but under "True Random" it wins 11% of the time. That means, if you EVER witness a 2-vs-2 where attacker rolls 1 die and wins, it is not balancing. Suggested test procedure: Start a game vs 2 bots. If not first, pause/restart until you are. You should get a couple 2-vs-2 shots per first turn. Just keep pausing/restarting until you win one. It will take 6 rolls on average. I've done it and won such a roll, proving discrete rolls are not balanced.
The "?" help for the ambiguously labeled "Dice Rolls" game parameter clearly says it adjusts how "blitz" rolls are executed. The term "blitz" means resolving battle in a single action, as opposed to a sequence of user controlled rolls. They use the term "Balanced Blitz" to refer to their scheme that skews blitz results in favor of the likeliest outcome, reducing/eliminating the chances of long shot outliers, and in their view, better rewarding strategic play. The term "True Random" means resolving blitz rolls as an unseen sequence of dice rolls without any "balancing". When you choose to roll 3, 2, or 1 dice (as opposed to blitz), you are choosing to roll according to the classic rules of the Risk board game, regardless of the setting of the "Dice Rolls" game parameter. Bottom line, under "Balanced Blitz", the GMs that choose discrete dice rolls rather than blitz when chipping against a stronger defender are correct, gaining a small improvement in results.
Interesting video. Thanks for posting.
This was super helpful! Thanks Slin :)
The manual dice do seem to be true random at least from my limited experience. I usually use them to weaken an army before blitzing them. You have a lot more games played though so I'll have to check that.
Well that's the thing it's hard to tell without lots of testing. It's only a 6% difference.
Theres like no chance 0
that I'm as unlucky with the manual dice as what I've been getting I lose soo much attacking and rarely get ahead when starting at an equal stack small or big numbers and blitz I get destroyed almost every time while attacking it's not realistic at all and this is like 350 games
Perfect argument against random crits too!!
Haha true
why yes i did quit comp tf2 to play risk full time how did you know
Lmao
I won my first game (very first game) after watching one of your vids last night... And I found the Die/Dice Rolls to be very straightforward & there was VERY little NUANCE to the ' a 6 versus a 2 or 3 almost always won'.
I just got 4th in this game this morning where EVERYONE except a starting A.I. (from a leaver) seemed to get very very very odd rolls in that consistently people were losing a 7 versus 2 or 8/9 versus 4 in the sense that '7-2 and the defender kept their territory (i.e. -6 troops for attacker and -1 for defender). It was happening again and again, it stopped 2 people from dying for about 3 turns even.
I noticed in the second game that 'True Random Rolls of the Die' was on and so I haven't been sure if that's the cause of the difference.
I do like True-Random though more I think in the end... The constant 'I know I'm going to crush an army completely and have enough to win by the time I get all the way across the map' is a bit much for me. I know each roll of a die in both modes can workout the same, so my losing 10 troops to an attacking army of 12 when they lost 1 troop is quite ridiculous even if not to the same degrees of ridiculousness that happens with your example of 20 losing to a 1.
Balanced Blitz affects a lot larger edge cases than you might realise. For example, the 4v1 has 100% on balanced, but its actual odds are only 91.15% (about 10 in 11). Considering how many players use a 4v1 to card trade most turns of the game, the certainty of success actually changes the flow of the game a lot.
I’ve about had it with the game. Feels like every time I’m attacking with a 20 against a 12 I somehow loose meanwhile also loosing while defending with a 100 against a 105… in balanced blitz.. makes zero sense. Great played games ending in the dumbest ways possible and consecutively, back to back.
In True Random. I've seen a 2 holding off an attack by a 9.
Not sure what the odds of that are...?
In general why would we want to not attack with our entire army? Should we not be going for kill shots?
To weaken a point and allow another attacker to kill
And it can limit troop transfers
I'm so fed up with risk. 2/6 people leave and the bot goes kamikaze. Cant really do much about it either.
This also happens way too much
I've been a victim of forgetting to ready up while playing on my phone or something (PC player)
@@1guard754 shame. I wish the game would allow you to make friends over steam so you can try to arrange a game with a bunch of players you know personally. Currently, you could only make risk friends which doesn’t facilitate communication.
@@anh-san well, there is a reason you can't play ranked with your friends
@@1guard754 I was more or less meaning a friends game but with the people you've already met and played with in random lobbies.
Sounds like balanced blitz uses a pseudo random distribution. In essence it would be that first throw is a true random, and the second throw lessens the chance for that same outcome.
In the 1 defending die case, first winning throw would have that 34,03 % chance, and second throw it would be less. How much less depends on the function..
The difference between 37 and 29 is 6?
Where is the blog link?
Thanks a lot
I don't see any links to the analysis
No offense but I think you have manual roll wrong. I've used it multiple times to make a bigger army smaller than mine then blitz it and win. I've even used it against capitals which have 3 dice defence
Right, but how do you know blitz roll wouldn't have had the same odds?
@@MRSLIN Good point. It just seems like from experience that it's more random than skewed in the attackers favor because there's also been plenty of times where I lose manual roll multiple times in a row
@@Daimxn A way to truly see if it is still a balanced blitz or has changed to a true random is to conduct a test of hundreds or thousands of rolls with the manual, then compare it with the brute force data from data genetics. If it closely matches(meaning it is statistically significant), then it is true random; conclude otherwise if not.
I like risk videos, but do you still do tf2 medic stuff?
ye
Not recently! I haven't been playing as much TF2 lately. Is that something you've been playing?
Is 15 vs 5 the same as 3 vs 1?
15 vs 5 has a slight advantage.
The blitz and manual feels the same ngl 🍌
Yeah like I'm saying, it's only 6% how can you even tell ya know?
You're wrong about 1V1 attacking. I can show you a big flaw in your theory
This should not be a thing because it should be RNG. Except it is not.. I can prove it with a recorded replay. 6 was hit 1 time by defense in 28 rolls..
The dice kill me. They talk about how the numbers are even but I play hundreds of games every month and 80% of the attacks that I engage in the other person kills me.