EPT[Idea] GEM: New LotV economy model - Page 21
| Forum Index > Legacy of the Void |
|
Penev
28528 Posts
| ||
|
ZenithM
France15952 Posts
| ||
|
Penev
28528 Posts
On July 03 2015 03:28 ZenithM wrote: Edit: Bah, nevermind. I chose not to as well, made a more random post instead. | ||
|
BlackLilium
Poland426 Posts
On July 03 2015 02:40 Geiko wrote: Ok bro, change that to "somewhere between 5 and 10% of games most likely" and let's continue the conversation from there. 5% - 10% games won due to base advantage? That doesn't sound so negative as you wanted to have in that sentence. Still, any number taken completely from hat is a manipulation. What you want to say is "not enough games..." - which is your opinion - but you hide it behind numbers, making it look as if it was a fact. Which is not. But even your 5% - 10% that is inaccurate. I am now looking at my notes when I was reviewing all the replays from the DH TL Open tournament. I see 28 games that I classified as "+1 base eco lead" (out of 135 I watched, which makes it 20.7%) and 8 games that I specifically classified as "3 base vs 4 base eco lead". Some of those were specifically noted as "3 base contain" - something you wouldn't see in HotS and would probably be problematic in LotV as well - because maintaining a contain requires higher income right now, not in the future. I picked 4 top games in that area to put in my replay analyzis. There are 4 because I have chosen that number of games, and not because there were no other games with required properties. In the end - let me stress that again - I don't mind if a number is inaccurate. I mind if the number is taken "from hat", as a "gut feeling", simply to disguise your opinion and treat it as a fact. | ||
|
Penev
28528 Posts
On July 03 2015 03:42 BlackLilium wrote: 5% - 10% games won due to base advantage? That doesn't sound so negative as you wanted to have in that sentence. Still, any number taken completely from hat is a manipulation. What you want to say is "not enough games..." - which is your opinion - but you hide it behind numbers, making it look as if it was a fact. Which is not. But even your 5% - 10% that is inaccurate. I am now looking at my notes when I was reviewing all the replays from the DH TL Open tournament. I see 28 games that I classified as "+1 base eco lead" (out of 135 I watched, which makes it 20.7%) and 8 games that I specifically classified as "3 base vs 4 base eco lead". Some of those were specifically noted as "3 base contain" - something you wouldn't see in HotS and would probably be problematic in LotV as well - because maintaining a contain requires higher income right now, not in the future. I picked 4 top games in that area to put in my replay analyzis. There are 4 because I have chosen that number of games, and not because there were no other games with required properties. In the end - let me stress that again - I don't mind if a number is inaccurate. I mind if the number is taken "from hat", as a "gut feeling", simply to disguise your opinion and treat it as a fact. Are there any concrete plans for a HMH TL tournament? | ||
|
Geiko
France1964 Posts
On July 03 2015 03:42 BlackLilium wrote: 5% - 10% games won due to base advantage? That doesn't sound so negative as you wanted to have in that sentence. Still, any number taken completely from hat is a manipulation. What you want to say is "not enough games..." - which is your opinion - but you hide it behind numbers, making it look as if it was a fact. Which is not. But even your 5% - 10% that is inaccurate. I am now looking at my notes when I was reviewing all the replays from the DH TL Open tournament. I see 28 games that I classified as "+1 base eco lead" (out of 135 I watched, which makes it 20.7%) and 8 games that I specifically classified as "3 base vs 4 base eco lead". Some of those were specifically noted as "3 base contain" - something you wouldn't see in HotS and would probably be problematic in LotV as well - because maintaining a contain requires higher income right now, not in the future. I picked 4 top games in that area to put in my replay analyzis. There are 4 because I have chosen that number of games, and not because there were no other games with required properties. In the end - let me stress that again - I don't mind if a number is inaccurate. I mind if the number is taken "from hat", as a "gut feeling", simply to disguise your opinion and treat it as a fact. My point was exactly that, a small number of games affected, not significant enough. Blizzard even said it. We’ve fully explored many of the community’s most popular models internally and took time to examine and evaluate the show matches as well. We watched the tournament matches, heard your responses, and we agree that the proposed change was not big enough compared to the Heart of the Swarm model. Now I appreciate you took the time to count the games etc. But first of all, you're analyzing your own model (or at least a model for which you have a significant bias), and second of all, you're using subjective criteria. What is "eco lead" ? Did you fix an arbitrary rate, where if player A gains x% income compared to standard HotS, then the game is classified as "eco lead" game ? I'm doing some nit-picking myself here, but just to say that the only things we can talk about are general tendencies. When I say 2% I mean, "few" , or "not significant enough". There's no manipulation, no one reads more into it than that tbh. | ||
|
Penev
28528 Posts
When I say 2% I mean, "few" , or "not significant enough". There's no manipulation, no one reads more into it than that tbh. I would agree but it shouldn't be a problem to adjust a little to the discussion partner if he wishes so. | ||
|
Geiko
France1964 Posts
On July 03 2015 02:08 BlackLilium wrote: Unless you counted those games - that's another manipulation from your side. Please either be precise with your numbers, or just say "not many". It's not synonymous. In the TLOpen DH tournament I have selected 4 where base advantage was most apparent, but there were more games. DH also helped stabilizing early aggression. You are also manipulating by implying that cheese is an inherent probelm of Double Harvesting idea. You very well know that it is not the case. Numbers causing the early aggression can be tuned down, and HMH 5-4 75% is an example of how it can be accomplished. Another thing is that any economic model that changes harvesting speed brings new timings (including cheese timings) and people need to learn when to scout to avoid them. I realize I forgot to address your second point. I never said cheese was a problem I just noted that one of the changes was that some cheeses are a bit more powerful. I didn't even say whether or not I thought that was a good thing. don't put words in my mouth. personally i couldn't care less, I like cheeses and with 12 worker start, they are getting a bit toned down anyways. it's not ineherant to the dh idea, but it is inherant to models with a non-linear incone curve early game. the curve is always convex (or concave I forget, f"(x)<0) so players staying on low worker count can exploit a slightly better economy than players saturating their bases. once again it's not a problem per se, it's just fact. | ||
|
Phaenoman
568 Posts
On July 03 2015 03:50 Geiko wrote: Now I appreciate you took the time to count the games etc. But first of all, you're analyzing your own model (or at least a model for which you have a significant bias), and second of all, you're using subjective criteria. What is "eco lead" ? Did you fix an arbitrary rate, where if player A gains x% income compared to standard HotS, then the game is classified as "eco lead" game ? I'm doing some nit-picking myself here, but just to say that the only things we can talk about are general tendencies. When I say 2% I mean, "few" , or "not significant enough". There's no manipulation, no one reads more into it than that tbh. Of course he is analyzing his "own" Modell. What else should he be doing? Analyzing urs without any data? It's primarily up to u to do that. But u like theorycrafting more apparently. Furthermore, stop this "I said this, but I meant that". Say what u really mean. That will help everyone here, including urself. | ||
|
Penev
28528 Posts
On July 03 2015 04:11 Phaenoman wrote: Of course he is analyzing his "own" Modell. What else should he be doing? Analyzing urs without any data? It's primarily up to u to do that. But u like theorycrafting more apparently. Furthermore, stop this "I said this, but I meant that". Say what u really mean. That will help everyone here, including urself. He means that Lilium is naturally biased. You're not helping btw | ||
|
Geiko
France1964 Posts
On July 03 2015 04:11 Phaenoman wrote: Of course he is analyzing his "own" Modell. What else should he be doing? Analyzing urs without any data? It's primarily up to u to do that. But u like theorycrafting more apparently. Furthermore, stop this "I said this, but I meant that". Say what u really mean. That will help everyone here, including urself. I'm not saying it's bad to analyze one's own model, I'm saying that if you do that, you should analyse it based on objective criteria. "I feel like the fact that this game was played on the DH mod has heavily influenced the outcome" is anything but objective. once again he did an amazing job going through all of these replays, but at this point, one can only speak about tendencies. which is what I'm doing. I feel like DH changes very little to the game as it is. it's a nice bonus but does it justifiy entirely changing the economy model for that ? | ||
|
BlackLilium
Poland426 Posts
On the other side, there is you, pulling numbers from a hat... | ||
|
Penev
28528 Posts
On July 03 2015 04:19 BlackLilium wrote: Of course I am biased about my own model. But I am trying to be as objective as possible in my opinion. Taking time, experimenting, doing measurements. On the other side, there is you, pulling numbers from a hat... Please don't, Geiko already responded to that. @Geiko: Pls use words instead of numbers if you don't mean numbers. | ||
|
Geiko
France1964 Posts
to get back on topic, I have a question for all of you. would you agree that GEM is currenty the only system that adresses (out of the main ones) the problem of maxing out too quickly and the overall feeling of being rushed to the late game ? why or ywhy not ? | ||
|
Penev
28528 Posts
On July 03 2015 04:26 Geiko wrote: Ok enough with the discussion on semantics. I feel like you guys are trying to get this locked again, and at this rate it certainly will because we're not having any interesting discussions... to get back on topic, I have a question for all of you. would you agree that GEM is currenty the only system that adresses (out of the main ones) the problem of maxing out too quickly and the overall feeling of being rushed to the late game ? why or ywhy not ? Well, HMH is adressing that too iirc? Worker inefficiency at the 9th without the initial DH boost I thought. Not sure which one does it best in that regard cause I'm watching HSC with beer. Well GEM would in practice | ||
|
Geiko
France1964 Posts
a good point for GEM is that it slows down the economy late game, without changing the early game. I believe it to be the inly system that currently achieves both these goals. | ||
|
RoomOfMush
1296 Posts
On July 03 2015 04:26 Geiko wrote:would you agree that GEM is currenty the only system that adresses (out of the main ones) the problem of maxing out too quickly and the overall feeling of being rushed to the late game ? why or ywhy not ? I dont agree. GEM might do it (depending on player behavior), but it is certainly not the only one. If players expand quickly enough GEM will not change the speed at which players max out. DH seems to slow down late game income pretty considerably. I could not test HMH yet but from the data I have seen it too seems to slow down the income pretty well. Comparing which one slows it down the most is beyond me at the moment. | ||
|
Penev
28528 Posts
On July 03 2015 04:39 Geiko wrote: well Hmh could in theory address it if the mining curves were adjusted to produce inferior economies with 16 workers compared to HOtS. Currently I believe they are being fitted to produce equivalent income to HotS with 16 workers, which does not slow down the economy late game (maybe lilium can confirm). a good point for GEM is that it slows down the economy late game, without changing the early game. I believe it to be the inly system that currently achieves both these goals. Make a graph  On July 03 2015 05:05 RoomOfMush wrote: I dont agree. GEM might do it (depending on player behavior), but it is certainly not the only one. If players expand quickly enough GEM will not change the speed at which players max out. DH seems to slow down late game income pretty considerably. I could not test HMH yet but from the data I have seen it too seems to slow down the income pretty well. Comparing which one slows it down the most is beyond me at the moment. DH has this early boost though (compared to HotS) | ||
|
BlackLilium
Poland426 Posts
On July 03 2015 04:39 Geiko wrote: well Hmh could in theory address it if the mining curves were adjusted to produce inferior economies with 16 workers compared to HOtS. Currently I believe they are being fitted to produce equivalent income to HotS with 16 workers, which does not slow down the economy late game (maybe lilium can confirm). HMH is fitted to match HotS income in 1-8 worker range. At 16 workers, HMH is at 75% of HotS. Earlier version of HMH was even slower and people actually complained about it! | ||
|
ChristianS
United States3304 Posts
-Testing, in addition to theorycrafting, is important in determining the efficacy of a model. -Of the models discussed, LotV and DH are most tested. HMH has some testing (I assume, since there's a mod for it, although I have no idea what any of the testing shows), and GEM is essentially untested. The next question, then, is whether or not GEM is promising enough in theory that we ought to dedicate some effort to trying to test it. Of the proposed models it is, on the surface, most similar to LotV. Upon closer inspection it might in some ways be the most different from HotS (in that it is hardest to keep HotS levels of income without increasing worker count). If the question is whether or not to test it, then Barrin's criticism is off-base, since it obviously makes no sense to say "we need test results to find out if we should test it." Edit: I realize that the use of quotation marks suggests that I am paraphrasing Barrin, which was not my intention. To the best of my understanding, Barrin was essentially saying the model can't really go any further without testing. I was merely pointing out that if the discussion is framed with the question "is this model good enough to test?" in mind, then that criticism no longer applies. Apologies for any misunderstandings. | ||
| ||
