|
gg ez discussion will not be continued in GD following this post |
On May 21 2015 10:56 cLutZ wrote:Show nested quote +On May 21 2015 10:28 Ketara wrote:On May 21 2015 10:24 Ansibled wrote:On May 21 2015 10:23 Caiada wrote:
Actual question, not argument.
Doesn't that just mean it's best to assess the situation (hope you can assess it right) and then make the best choice? Basically just arguing against deciding randomly?... I think the point is that in situations where the mid and jungler disagree on who should get the buff, then it should default to the midlaner. Cheep is arguing that it should default to the midlaner, the rest of us are arguing that it should default to the jungler making an educated decision about who should get it based on the circumstances of that individual game, which will likely lead him to giving it to the mid laner in the majority of circumstances anyway. Cheeps argument is that this doesn't make sense because mid and jungle should have equal game information and equal skill levels, so this doesn't make sense because if they don't agree its impossible to tell who is right. But that's not the case, because while players should be perfect, they aren't, and in reality the jungler has easier access to game information than the mid laner does, so he should make the decision and the mid laner should trust that decision. This is what all the diamonds and masters players are saying and cheep basically refuses to admit that he doesn't understand the game as well as Scip and should just concede this point. At least that's my following of it. I don't really know what the fuck is going on. Your position makes no sense because its basically you saying it defaults to the jungle making an informed choice, but we only care in the case that said choice is one that the mid disagrees with (aka him keeping blue). We only care about this situation, and what is "correct" then. Teut thinks this is 1% or so, it seems to me to be much higher, probably 20%. Your solution is actually to just give it to the jungle. The underlying premise to the whole argument is that junglers take blue too often, which, if true, means your argument that they know better is wrong. So it all boils down to an empirical argument (which people are constantly confounding with random situations involving yasuo), of if the current mid transfer percentage is too low, and if so, the tie should go to the mid.
I have been in situations as jungler before where I want mid to have blue and he refuses to take it. I've also been in situations blue side where I give it to top lane instead of mid lane. So a disagreement doesn't necessarily always mean the jungler takes it for himself.
|
You're introducing situations outside of the decision point then. I will say that there is an alternate hypothesis, we would call it the "greedy mid" hypo, where junglers should have the deciding say, because mid asks for blue at a much higher rate than its ideal.
|
Also, my argument isn't that the jungler knows better, its that the jungler is better equipped to know better, so if there is a debate the jungler with his superior information should just be listened to.
Players obviously make incorrect decisions. That doesn't mean they shouldn't be trying to make those decisions.
|
Czech Republic11293 Posts
On May 21 2015 11:01 JimmiC wrote:Show nested quote +On May 21 2015 10:57 Scip wrote:On May 21 2015 10:55 JimmiC wrote:On May 21 2015 10:48 Scip wrote:On May 21 2015 10:34 JimmiC wrote:
Rest of us is untrue. I agree that it should go to mid if there is a disagreement. I also think that in the end jungler will get when he wants because of smite.
Ketera I find it interesting that u think jungle should get choice but in your games try to steal it with Lux ult. Do as I say not as I do? I don't understand if there's a disagreement between you (the jungler) and the middle lane about who should get blue then you clearly think you should get the blue if you think you should get the blue but think that your middle laner is more likely to be right then you're just disagreeing with yourself wtf? If you're an outside spectator you USUALLY have more info to go off of than just the player's position anyway so whatever. I was trying to say I think it should go to the mid, but the jungler will take it when he/she wants because of smite. Your last statement could be true, but I believe that having a non ragey mid is probably more valuable then the small boost you would get in the jungle. Could even give the mid a psychological boost. But sometimes it SHOULDN'T go to the mid wat do then That non-ragey mid can be more valuable depends and it's up to your personal experience to decide how much of a concern it is. In my experience it's not a problem at all, but maybe things are a bit different on your server/level of play. I think this shows there is non hard and fast rule and you have to make the judgement. Like I said earlier if I think I should take it as jungler I'll state that and why and usually no big complaint. If the guy says he needs it and blah blah I usually just give it to him for reasons stated.
Yeah, well, depends on your mentality then. My general policy is to not do things that I think are wrong, but I can see why you play it this way.
|
On May 21 2015 10:55 Ketara wrote:
Basically
The entire argument settles around this hypothetical disagreement between the jungler and mid over who should get blue buff.
And it assumes that the opinions of these two parties have equal value.
That assumption is wrong. The junglers opinion is more important.
Only if the jungler is an unbiased decision maker. I don't think he is.
So in a less dumb conversation lets do that thing with the Ashe Rune Pages i said i would do earlier
@Alaric
On May 21 2015 01:37 Goumindong wrote:Show nested quote +On May 20 2015 17:03 Alaric wrote:On May 20 2015 10:33 Goumindong wrote:On May 20 2015 10:25 Sonnington wrote:
I haven't tested or played enough Ashe to notice, but does shiv crit every single time she auto's an enemy that already has her passive? I'd imagine it would since all her autos are minicrits after her slow is applied. It does yes. But its the same multiplier as all of your attacks. So if you have IE and Shiv the proc hits for 170 damage (so long as you don't accidentally waste it on a non-slowed target and you don't have any other crit chance but IE/Shiv). Also worth noting that you should not ever blind pick NuAshe because Olaf and Yi are immune to her crits when they ult. They said Shiv uses your crit% stat, not the auto crit. I don't understand what this means. Like if you have just shiv does it do 120 or 130 or 100/200 80/20%? When I tested on the PBE I thought it was 130. It could be 120 but i am near certain it wasn't 100/200 80/20. If it is 120 (and so 160 or 140 with IE and shiv) then it's holy goodness an even worse buy than I thought. Aside: when I get home tonight I will spreadsheet up some Ashe rune options to see what cN be done. And also do some testing on shiv damage.
So i went and spreadsheeted up some ashe pages. Well, ok, not that many.
Some things to note. These are the efficiency numbers of crit/as/ad runes. AD of equivalent color are the base. With the AD efficiency its just magnitude compared to blue AD runes. Higher is better, and so if you're maximizing you use those runes first(over AD) if you're going to have those runes and/or AD in that slot.
Red Quint Yellow Blue
AS/AD 1.789473684 1.956521739 1.813953488 2.285714286
Crit/AD 0.978947368 0.808695652 0.976744186 1
AD/Blue 3.392857143 8.214285714 1.5 1
So i decided to check the best tradeoff for a rune page which would be 9 AD Reds, 3 AS Quints, Who Cares Blues/Yellows.
Trading off 1 AD for 1 Crit red at a time. The critical strike/damage value for Ashe does not break even until you have a BF sword. When it does the overall Damage increase on crit auto attacks increase of having 8 crit reds and 1 AD is .3%.
When you get your IE, then critical strike is strictly better than AD to the tune of 3% and this will increase as you continue to stack AD from other sources but decrease as you continue to stack crit (in the end, AD is easier to stack than crit, so you should wind up with more lategame DPS using crit runes). When you're just starting out, you lose nearly 4.5 damage per attack however, which is 6% of your auto attack DPS, plus a good chunk off your volley and non-crit attack.
At level 18 full build (IE,PD, BT,LW,Scimitar, one dragon) it makes about 4.5% difference. With a second PD instead of a scimitar it makes about 2.6% difference [Red Pot/Baron have basically zero effect on this]
Here are the raw values its just the damage value because AS is constant across levels. 0 = Zero crit reds, 9 = 9 crit reds
Crit Red Level 1 Level 2 Level 3 Level 4 Level 5 Level 6 Level 7 Level 8 Level 9 Level 10 6+Pick 6+BF 10+IE
0 79.3155 82.15075 84.986 88.92125 92.39175 95.227 99.16225 103.0975 105.93275 109.868 124.102 147.202 257.432
1 78.87954975 81.7387705 84.59799125 88.566512 92.0663535 94.92557425 98.894095 102.8626158 105.7218365 109.6903573 124.0446993 147.3399993 258.5867109
2 78.425046 81.3082375 84.191429 88.1932205 91.7224035 94.605595 98.6073865 102.609178 105.4923695 109.494161 123.968845 147.459445 259.7135915
3 77.95198875 80.859151 83.76631325 87.8013755 91.3599 94.26706225 98.3021245 102.3371868 105.244349 109.2794113 123.8744373 147.5603373 260.8126419
4 77.460378 80.391511 83.322644 87.390977 90.978843 93.909976 97.978309 102.046642 104.977775 109.046108 123.761476 147.642676 261.883862
5 76.95021375 79.9053175 82.86042125 86.962025 90.5792325 93.53433625 97.63594 101.7375438 104.6926475 108.7942513 123.6299613 147.7064613 262.9272519
6 76.421496 79.4005705 82.379645 86.5145195 90.1610685 93.140143 97.2750175 101.409892 104.3889665 108.523841 123.479893 147.751693 263.9428115
7 75.87422475 78.87727 81.88031525 86.0484605 89.724351 92.72739625 96.8955415 101.0636868 104.066732 111.721439 123.3112713 147.7783713 269.4122585
8 75.3084 78.335416 81.362432 85.563848 89.26908 92.296096 96.497512 100.698928 103.725944 107.92736 123.124096 147.786496 265.89044
9 74.72402175 77.7750085 80.82599525 85.060682 88.7952555 91.84624225 96.080929 100.3156158 103.3666025 107.6012893 122.9183673 147.7760673 266.8225089
So it looks like you may want to drop some AS due to the early free AS from the crit mastery. But you're almost assuredly going to want to trade them for AD/pen runes rather than crit runes. The lategame DPS increase probably isn't worth missing a few CS/trading poorly/being a few hundred gold behind and i feel like the magnitude of the early DPS loss is significant enough to do that or more. This is especially important considering how snowbally NuAshe's laning phase potentially is.
The increase in DPS from a red pot (when you have IE, PD, Whisper, BT, and Scimitar) is 6.7% to put this in perspective.
|
|
|
this is why I play the island role
|
Are we still debating blue buff? Cheep is a true monster
|
On May 21 2015 10:06 Goumindong wrote:Show nested quote +On May 21 2015 08:35 Fildun wrote:On May 21 2015 08:28 Goumindong wrote:On May 21 2015 08:01 Fildun wrote:
because the 70/30 split doesn't say anything about the chance of someone being wrong in a single circumstance. It's basic stochastics.
Uhh yes it does. It says that if you choose the 70% choice every time you will be wrong 30% of the time and if you choose the 30% choice you will be wrong 70% of the time. The 70/30 split., in this instance, defines the individual probabilities. (Though technically we say that the split is defined by the probabilities and not the other way around this has no effect on the predicted outcomes, which is to say "the chance of someone being wrong" The problem is that you're seeing the chance of someone being wrong as the same thing as the chance of a mid getting blue buff. Actually, I feel like we've defined wrong differently. No, not at all. Suppose we are in a situation where its giving blue buff to the mid laner has more beneficial outcomes compared to giving it to the jungler 60% of the time. Should we give the blue buff to mid every time? A: Yes. We should, just like you should take a 60% 1 to 1 bet every time. It doesn't matter if its offered to you once or a hundred times you should take it every time**. This is especially true since we can invert the situation, going with the jungler then is a 40% bet and you would be stupid to take a 40% bet. Ever, once, you should not take that bet. The question is "is it really better to give it to the mid laner" and the answer is "yes in the majority of times it is, there are edge cases where its not*, but for the most part you want to donate if you can, because its better on the mid laner." And, just like you should take favorable bets every time, you should give the buff to the mid laner So then we get to "well does the jungler know better than the mid" and well maybe the answer is yes, (it should be yes) but junglers don't seem to make the right decision since the overall donation rate seems to be lower than it ought to be(though this is just a casual observation and not recorded data). *manaless, non-ult dependent mid laners and/or with low ult CD's. And/or someone so far behind you've lost the game anyway unless your ult dependent jungler can utilize the CDR to gank on cooldown. And the answer, just like when betting on a 60% bet is to bet on mid. *provided you're liquid enough, but this type of risk aversion doesn't come into play in league
Nonononono. You take the bet 60% of the time, because you try to assess whether the situation is in the 60% or in the 40%. That's the whole idea.
Let's assume that if mid and jungler agree it's the right call.
The problem is that there is a 1% somewhere where the mid and jungler disagree.
Let's assume that the 60/40 split is the optimum, and there is 1% of disagreement. We don't know whether 59.9% already goes to the mid or 59.1% already goes to the mid. In the rirst case, in the case of a disagreement, the jungler is right 90% of the time and in the second situation he is right 10% of the time. This is only when there is a disagreement, but that's what the whole argument was about.
|
I don't get how this is such a complicated topic.
X% of the time (lets say 55%) jungler and mid agree it should go to mid. Y% (lets say 35%) jungler and mid agree it should go to jungler. The remaining 10% of the time are the fringe cases where it's a hard judgement call, maybe mids losing, maybe he doesn't "need" it on his champ, maybe jungle needs it for 6 etc etc.
In those cases mid and jungle might make different value judgements and disagree on what the best course of action to win the game is. In those cases, both players do the play they think will win the game, and jungler has this magical ability called "fuckyoui'majunglerwhohassmite" and the jungler gets blue.
SoloQ ranks people based on their ability to make good decisions, if you as a jungler think that taking blue is better for winning then giving it to mid, you take it. If you as a mid laner think you getting blue is better for winning, tough shit, you don't have smite, so unless you saw him use it and want to steal, suck it up.
It's really not that complicated, neither persons "opinion" is more important, the jungler just has the final say because he has smite. End of.
|
On May 21 2015 17:38 Fildun wrote:Show nested quote +On May 21 2015 10:06 Goumindong wrote:On May 21 2015 08:35 Fildun wrote:On May 21 2015 08:28 Goumindong wrote:On May 21 2015 08:01 Fildun wrote:
because the 70/30 split doesn't say anything about the chance of someone being wrong in a single circumstance. It's basic stochastics.
Uhh yes it does. It says that if you choose the 70% choice every time you will be wrong 30% of the time and if you choose the 30% choice you will be wrong 70% of the time. The 70/30 split., in this instance, defines the individual probabilities. (Though technically we say that the split is defined by the probabilities and not the other way around this has no effect on the predicted outcomes, which is to say "the chance of someone being wrong" The problem is that you're seeing the chance of someone being wrong as the same thing as the chance of a mid getting blue buff. Actually, I feel like we've defined wrong differently. No, not at all. Suppose we are in a situation where its giving blue buff to the mid laner has more beneficial outcomes compared to giving it to the jungler 60% of the time. Should we give the blue buff to mid every time? A: Yes. We should, just like you should take a 60% 1 to 1 bet every time. It doesn't matter if its offered to you once or a hundred times you should take it every time**. This is especially true since we can invert the situation, going with the jungler then is a 40% bet and you would be stupid to take a 40% bet. Ever, once, you should not take that bet. The question is "is it really better to give it to the mid laner" and the answer is "yes in the majority of times it is, there are edge cases where its not*, but for the most part you want to donate if you can, because its better on the mid laner." And, just like you should take favorable bets every time, you should give the buff to the mid laner So then we get to "well does the jungler know better than the mid" and well maybe the answer is yes, (it should be yes) but junglers don't seem to make the right decision since the overall donation rate seems to be lower than it ought to be(though this is just a casual observation and not recorded data). *manaless, non-ult dependent mid laners and/or with low ult CD's. And/or someone so far behind you've lost the game anyway unless your ult dependent jungler can utilize the CDR to gank on cooldown. And the answer, just like when betting on a 60% bet is to bet on mid. *provided you're liquid enough, but this type of risk aversion doesn't come into play in league Nonononono. You take the bet 60% of the time, because you try to assess whether the situation is in the 60% or in the 40%. That's the whole idea. Let's assume that if mid and jungler agree it's the right call. The problem is that there is a 1% somewhere where the mid and jungler disagree. Let's assume that the 60/40 split is the optimum, and there is 1% of disagreement. We don't know whether 59.9% already goes to the mid or 59.1% already goes to the mid. In the rirst case, in the case of a disagreement, the jungler is right 90% of the time and in the second situation he is right 10% of the time. This is only when there is a disagreement, but that's what the whole argument was about.
Is it really so hard to realize that you cannot predict the future and just let it go
|
Yes, because the goal of the whole game is to predict the future.
Also, I don't see how your comment is constructive in any way to our argument.
|
You (Killerdog) make it less complicating by ignoring the interesting question. What is the rational action for the jungler? Should the midlaner's testimony be simply ignored, if we think the midlaner is an epistemic peer?
One strategy is to argue that the midlaner is in fact not an epistemic peer (e.g. jungler is in a better position to tell the gamestate, or the jungler has additional higher-order evidence about his/her reasoning that he/she does not have for the midlaner) -> rational for the jungler to stick with his/her decision.
The other strategy is to argue that the midlaner is in fact an epistemic peer -> the jungler needs to revise his/her course of action. It is very unclear how to revise the intention though, because simply following the midlaner would also be irrational due to ignoring the jungler's reasoning.
|
On May 21 2015 17:59 Goumindong wrote:Show nested quote +On May 21 2015 17:38 Fildun wrote:On May 21 2015 10:06 Goumindong wrote:On May 21 2015 08:35 Fildun wrote:On May 21 2015 08:28 Goumindong wrote:On May 21 2015 08:01 Fildun wrote:
because the 70/30 split doesn't say anything about the chance of someone being wrong in a single circumstance. It's basic stochastics.
Uhh yes it does. It says that if you choose the 70% choice every time you will be wrong 30% of the time and if you choose the 30% choice you will be wrong 70% of the time. The 70/30 split., in this instance, defines the individual probabilities. (Though technically we say that the split is defined by the probabilities and not the other way around this has no effect on the predicted outcomes, which is to say "the chance of someone being wrong" The problem is that you're seeing the chance of someone being wrong as the same thing as the chance of a mid getting blue buff. Actually, I feel like we've defined wrong differently. No, not at all. Suppose we are in a situation where its giving blue buff to the mid laner has more beneficial outcomes compared to giving it to the jungler 60% of the time. Should we give the blue buff to mid every time? A: Yes. We should, just like you should take a 60% 1 to 1 bet every time. It doesn't matter if its offered to you once or a hundred times you should take it every time**. This is especially true since we can invert the situation, going with the jungler then is a 40% bet and you would be stupid to take a 40% bet. Ever, once, you should not take that bet. The question is "is it really better to give it to the mid laner" and the answer is "yes in the majority of times it is, there are edge cases where its not*, but for the most part you want to donate if you can, because its better on the mid laner." And, just like you should take favorable bets every time, you should give the buff to the mid laner So then we get to "well does the jungler know better than the mid" and well maybe the answer is yes, (it should be yes) but junglers don't seem to make the right decision since the overall donation rate seems to be lower than it ought to be(though this is just a casual observation and not recorded data). *manaless, non-ult dependent mid laners and/or with low ult CD's. And/or someone so far behind you've lost the game anyway unless your ult dependent jungler can utilize the CDR to gank on cooldown. And the answer, just like when betting on a 60% bet is to bet on mid. *provided you're liquid enough, but this type of risk aversion doesn't come into play in league Nonononono. You take the bet 60% of the time, because you try to assess whether the situation is in the 60% or in the 40%. That's the whole idea. Let's assume that if mid and jungler agree it's the right call. The problem is that there is a 1% somewhere where the mid and jungler disagree. Let's assume that the 60/40 split is the optimum, and there is 1% of disagreement. We don't know whether 59.9% already goes to the mid or 59.1% already goes to the mid. In the rirst case, in the case of a disagreement, the jungler is right 90% of the time and in the second situation he is right 10% of the time. This is only when there is a disagreement, but that's what the whole argument was about. Is it really so hard to realize that you cannot predict the future and just let it go
The entire point of league is to attempt to predict the future. That's called decision making.
You take into account all the things you know about the situation, and use your base knowledge of the game and experience to try and predict how the game plays out, and based off that you make decisions to favourably alter that flow.
I mean I guess you could literally flip a coin at every single decision fork, because "the future is unknowable!!!" but against all odds, I predict that flashing under their tower at level 1 might kill me.
+ Show Spoiler +
On May 21 2015 18:05 Prog wrote:
You (Killerdog) make it less complicating by ignoring the interesting question. What is the rational action for the jungler? Should the midlaner's testimony be simply ignored, if we think the midlaner is an epistemic peer?
One strategy is to argue that the midlaner is in fact not an epistemic peer (e.g. jungler is in a better position to tell the gamestate, or the jungler has additional higher-order evidence about his reasoning that he/she does not have for the midlaner) -> rational for the jungler to stick with his/her decision.
The other strategy is to argue that the midlaner is in fact an epistemic peer -> the jungler needs to revise his course of action. It is very unclear how to revise the intention though, because simply following the midlaner would also be irrational due to ignoring the jungler's reasoning.
That's a fundamental question as to whether you trust your own judgement in a game. If your jungler is obviously better then you, or if your mid laner is obviously better then you, then you might listen to them. I (generally) won't try and steal a buff from scip if he actually has a logical reason for wanting it. And in the same vein, if i'm jungling and a mid laner makes an actual argument for needing blue, that, after thinking through, makes sense to me and changes my position, then I'd give him the buff. But this is all information you take into account in deciding whether taking or giving blue is the best course towards winning.
If, after taking all these things into account, my analysis is still that me taking it is better, then I should do what I think is right. Either it's the right decision and it goes well, or it's the wrong decision and you learn from that and grow as a player, and using that knowledge make a more informed decision next time.
If you think something through all the way, taking everything into account, then do the opposite action to what you think is right, then I don't know what to tell you :S
|
The cases of x being obviously better are ruled out by stating that the jungler thinks the midlaner is an epistemic peer, that is, the midlaner appears to have the same evidential basis and the same skill to make a judgment. Now what is the rational thing to do?
If you simply ignore your midlaner you are irrational, because you ignore evidence (the midlaner's testimony) that is equally well supported as your own. If you do what the midlaner says you also ignore evidence (your own this time), so you are irrational as well. And there seems to be no middle ground possible.
So either you think that jungler and midlaner are no actual peers (in a non-obvious way), or you opt for some middle ground option that I cannot see (if its more than 1 blue buff situations you can split perhaps).
|
It's basically assisting enemy team if you give blue to mid who's just going to feed it to the opponent. It can also be what the mid needs to be able to farm safely if they in fact are a good player who just got unlucky and needs that little advantage / equalizer to get back into the game properly. In my opinion, it's down to your judgement and you can't just generalize it. An awful player might get lucky and be 2-0, an amazing player might get unlucky or make the very rare bad play and be 0-2.
|
If the mid and jungler are epistemic peers, shouldn't they always come to the same conclusion, because they both have the same information available?
|
|
|
On May 21 2015 18:21 Prog wrote:
The cases of x being obviously better are ruled out by stating that the jungler thinks the midlaner is an epistemic peer, that is, the midlaner appears to have the same evidential basis and the same skill to make a judgment. Now what is the rational thing to do?
If you simply ignore your midlaner you are irrational, because you ignore evidence (the midlaner's testimony) that is equally well supported as your own. If you do what the midlaner says you also ignore evidence (your own this time), so you are irrational as well. And there seems to be no middle ground possible.
So either you think that jungler and midlaner are no actual peers (in a non-obvious way), or you opt for some middle ground option that I cannot see (if its more than 1 blue buff situations you can split perhaps).
Soloq gives each player X decisions they get to make to try and carry the game. Your goal as a player in soloq is to use your split of the teams decisions as optimally as possible to try and carry the game to a victory. Why would you, as a jungler, cede one of your decisions to midlane?
I'm not saying the mid lane asking for blue shouldn't factor into your equation when determining what the best course of action is, of course if they ask for it, that adds evidence in favour of sharing it, but if, after taking that into account, I still feel that I should take it, then I should.
I feel like you're just playing devils advocate, but don't you agree that the goal of soloq is to make as many decisions you feel are "correct" as possible? You can't go into soloq and be afraid to make any decisions because you think everyone in the game is better then you, and expect to do well.
Assuming you play all roles, then 50% of the time you'll be on the mid laner side of the equation, and 50% of the time you'll be on the jungler side. Now if you want to have as many chances to carry games as possible, you want to be allowed as many important decision forks as possible. Therefore, I'd propose that, if you trust in your ability to make decisions, you should play in such a way that you get to make as many decisions as often as possible. That is, make the call in 100% of the games you're jungling, and every game you're mid laning where the jungler lets you make the call.
If your ideology is that everyone in your average game is actually "smarter" then you, and you defer to your teammates opinions on everything, even when you objectively think they're wrong, then I think we've found an ideological split that makes this conversation pointless.
|
Kdog you wot mate. If the midlaner is likely to make a better decision than you you let him take the decision. Ofcourse this depends on the game, but hoarding all the decisions only works when you're smurfing.
Also I don't see how Prog even said half the things you're trying to reply to.
|
|
|
|
|
|
EPT
