|
I know most people tend to either stop probe production entirely after a point (4gate allin) or build constantly to ensure there are plenty of probes to transfer when an expansion does go up, but assuming you want to expand, is that in fact the best way?
Doesn't it make more sense to build a certain number of probes (somewhere between 2-3 per patch I'd assume), then cut probes entirely until you lay down a nexus, and then start chronoing them out again?
Think of two protoss players: assuming their spending on other things is identical, there will be a point in time where one player will be able to build a nexus, while the other will have 6-7 more probes but no nexus (6-7 instead of 8 because a probe costs 62.5 when you account for supply). Then the nexus-first guy will build more probes to make up for cutting, and the always-probe guy will continue to build probes and, at some later point, a nexus.
So the always-probe guy makes better use of his first nexus' probe-building potential, but has to spend early minerals on probes beyond saturation that give no benefit until a new nexus is built.
The probe-cut guy wastes some probe-constructing time, but gets a second nexus faster, which gives: a second probe-producing building sooner, more total chronoboost energy, and allows workers to mine at unsaturated patches sooner.
As an added bonus, the guy who cuts probes hasn't locked in as much spending on minerals until the point where he builds his nexus, giving him a little extra flexibility to deal with any quirky enemy builds he manages to scout. And if the issue is 'you can't make a nexus until you can defend it', then cutting probes allows you to get more defenses faster, meaning you can lay down the expo sooner.
This logic would apply to 3rd and 4th bases, as well.
Am I missing anything? It looks to me like cutting probes, then resuming only after placing a nexus, is better. The only question, it seems, is when to cut. @2 probes/mineral patch (max efficiency)? @ 2.5/patch (less efficient but higher total output)? @ 3/patch (maximum saturation)?
I've been trying cutting at 2 probes/patch lately, and am happy with the results, but I'm open to opinions.
Oh and the same idea might apply to terran (faster MULE is nice) but seems meaningless to zerg as they always expand faster than they can saturate anyways =p
EDIT: can't seem to add the [D] tag to title via edit, sorry I forgot that
|
It's a nice idea, but it would be nice if you did the math before posting so that your OP actually has content instead of speculation and guessing.
In real terms though, I don't think cutting probes to make a nexus is ever better. Cutting units is almost always the better option.
|
|
|
worker transfer means your nexus comes up when i have 8 probes to trnasfer without desaturating my main. cutting probes for better probe production is silly
|
also, even if you cut units, i would argue that the exact same logic still applies. If protoss (a) and protoss (b) both cut units, but one also cuts probes, he will get a faster nexus in exchange for less probes.
|
On April 22 2011 15:06 Teivospy wrote:
worker transfer means your nexus comes up when i have 8 probes to trnasfer without desaturating my main. cutting probes for better probe production is silly
it's not for better probe production, it's for more efficient mining via unused mineral patches. If you have 8 extra probes to transfer without desaturating, then that means you spent 400 + supply on 8 probes that do absolutely nothing until your expansion goes down.
|
On April 22 2011 15:05 Keilah wrote:
what math could I do?
not hard calculations necessarily but you can always create custom game and time yourself using both strategies see where the advantages lie if any.
this is just mostly guesswork/speculation. not very accurate i think. do you at least have replays of this getting you ahead?
|
On April 22 2011 15:11 Keilah wrote:Show nested quote +On April 22 2011 15:06 Teivospy wrote:
worker transfer means your nexus comes up when i have 8 probes to trnasfer without desaturating my main. cutting probes for better probe production is silly it's not for better probe production, it's for more efficient mining via unused mineral patches. If you have 8 extra probes to transfer without desaturating, then that means you spent 400 + supply on 8 probes that do absolutely nothing until your expansion goes down.
except in starcraft 2 saturation comes at 26+ probes including gas. so unless you have over 26 probes you're going to have exactly the same mining, except you're just going to have less probes because you cut them for a nexus.
|
On April 22 2011 15:14 Mykill wrote:not hard calculations necessarily but you can always create custom game and time yourself using both strategies see where the advantages lie if any. this is just mostly guesswork/speculation. not very accurate i think. do you at least have replays of this getting you ahead?
I sort of tried that, and to keep the data pure i built nothing but probes, pylons, and a nexus, but I realized that if I do that, i get enough minerals to make a new nexus much faster than i can reach 2 probes/patch, so nonstop production never has a chance to deviate from cutting at 2/patch.
If I start building gateways and gas and units and whatnot, there's so much room for error where i just fuck up a little. I suppose that if I did some standard build and played near-flawlessly vs the very easy AI I could try both techniques and look at them in the replay. I'd have to be very careful to build the exact same buildings and units (other than a pylon or two), but it could be done.
|
On April 22 2011 15:18 rauk wrote:Show nested quote +On April 22 2011 15:11 Keilah wrote:On April 22 2011 15:06 Teivospy wrote:
worker transfer means your nexus comes up when i have 8 probes to trnasfer without desaturating my main. cutting probes for better probe production is silly it's not for better probe production, it's for more efficient mining via unused mineral patches. If you have 8 extra probes to transfer without desaturating, then that means you spent 400 + supply on 8 probes that do absolutely nothing until your expansion goes down. except in starcraft 2 saturation comes at 26+ probes including gas. so unless you have over 26 probes you're going to have exactly the same mining, except you're just going to have less probes because you cut them for a nexus.
Is that a fact? I was under the impression that 3/patch was absolute max saturation, but that minerals/probe/minute decreased for every probe beyond 2/patch.
Anyways, 26 seems like a reasonable number to stop at. I'm pretty sure making 26 probes, then a nexus, then resuming probes would be more efficient than, say, making 32 probes, then a nexus, and continuing to make probes.
|
On April 22 2011 15:20 Keilah wrote:Show nested quote +On April 22 2011 15:14 Mykill wrote:On April 22 2011 15:05 Keilah wrote:
what math could I do? not hard calculations necessarily but you can always create custom game and time yourself using both strategies see where the advantages lie if any. this is just mostly guesswork/speculation. not very accurate i think. do you at least have replays of this getting you ahead? I sort of tried that, and to keep the data pure i built nothing but probes, pylons, and a nexus, but I realized that if I do that, i get enough minerals to make a new nexus much faster than i can reach 2 probes/patch, so nonstop production never has a chance to deviate from cutting at 2/patch. If I start building gateways and gas and units and whatnot, there's so much room for error where i just fuck up a little. I suppose that if I did some standard build and played near-flawlessly vs the very easy AI I could try both techniques and look at them in the replay. I'd have to be very careful to build the exact same buildings and units (other than a pylon or two), but it could be done.
Just do a build, 2 gates, 2 gasses fully saturated and then constantly produce zealots, then one build you can cut probes and the other you not cut probes.
|
Depends on your gameplan/situation. By all means, if you have 26 probes your best-case scenario would be to throw down a nexus as quickly as possible, but if your build requires you to spend the 400 minerals elsewhere, or if you scout your enemy and he 3 raxed and is running up to your base, then it would be wise to not throw down the nexus yet (until you get a couple immortals and a decent army mix).So in this situation, you might as well continue producing probes, even if you oversaturate; the extra probes you make now are probes you don't have to make later when your expo finishes.
If you see that your enemy cannot really threaten you, and if your gameplan allows you to expand, then yeah, throw it down.
|
Interesting idea, I assume you would spend chrono on units/upgrades until the next nexus went up? Or would you store chrono for the probes once the expo goes up? Keep in mind that having those extra probes rdy, right when the expo pops is an advantage in itself for the moment and may give you a long term momentum swing, especially once both parties are on 3+ bases. This means he'll always have more probes than you, which is usually the way one tells who is ahead. But you're right you gain a lot off of spending your money on units when the probes that you would produce won't be helping much in the now.
Another interesting thing to note is that this is why fast expansions are so great. No over saturation/decreased over saturation. You ask, should I stop building probes because they will be useless for a small period? You can side step the issue by building probes non stop and expanding before it ever becomes an issue.
|
When you're saturated on one base, rally your nexus to your natural's minerals and distance mine.
|
On April 22 2011 15:27 Hierarch wrote:Show nested quote +On April 22 2011 15:20 Keilah wrote:On April 22 2011 15:14 Mykill wrote:On April 22 2011 15:05 Keilah wrote:
what math could I do? not hard calculations necessarily but you can always create custom game and time yourself using both strategies see where the advantages lie if any. this is just mostly guesswork/speculation. not very accurate i think. do you at least have replays of this getting you ahead? I sort of tried that, and to keep the data pure i built nothing but probes, pylons, and a nexus, but I realized that if I do that, i get enough minerals to make a new nexus much faster than i can reach 2 probes/patch, so nonstop production never has a chance to deviate from cutting at 2/patch. If I start building gateways and gas and units and whatnot, there's so much room for error where i just fuck up a little. I suppose that if I did some standard build and played near-flawlessly vs the very easy AI I could try both techniques and look at them in the replay. I'd have to be very careful to build the exact same buildings and units (other than a pylon or two), but it could be done. Just do a build, 2 gates, 2 gasses fully saturated and then constantly produce zealots, then one build you can cut probes and the other you not cut probes.
OK so I did this. I did a standard economic build, no scouting, gate on 13, gas on 15, 2nd gas after building 19th probe, constant zlot production, build 2nd gate as soon as I have 150 minerals without cutting either probes or zlots, then pump probes and zealots with pylons nonstop and build the expansion nexus asap without cutting anything. I used all chronoboosts on probes, as soon as possible, and played pretty cleanly with few 'could have saved a half second here and there' moments in each game.
One game I never stopped making probes. At 10:08 I had:
-21 zealots (19 with 2 more building)
-10 pylons
-54 probes
-96/100 food
-1310 minerals
-1632 gas
The next game, I built probes until I had 17 on minerals, then cut until I got a nexus up, then resumed. This resulted in me storing surprisingly little chronoboost, I think I was on 35 energy or so when I restarted probes. At 10:08 i had:
-21 zealots (19 with 2 more building)
-11 pylons
-60 probes
-102/108 food
-1560 minerals
-1648 gas
A net gain of 6 probes, 1 pylon, and 250 banked minerals, totaling 650 extra minerals by 10:08 for the game where I cut probes at 17 on minerals. Pretty nice!
A few notes:
-i built the geyser a tiny bit earlier in the 2nd game by accident, hence more gas at 10:08
-continuing to chronoboost probes after saturated seemed... retarded and unrealistic. It would probably be closer if I just stored the chrono for the probe-maker once he was saturated.
-I didn't take a 3rd in either game, but if I had again stopped producing probes at 18 on minerals/base, I would have been able to build another nexus MUCH faster in the second game, presumably with the result that I would be even farther ahead if I'd expanded again.
-In both games, I had 2 bases approximately saturated at 10:08
-I meant to stop at 10:00 but messed up a little, apparently if you save you can't load the replay so I was forced to manually count
-because of my mistake I have no replays. But feel free to try it yourself!
|
On April 22 2011 17:02 Keilah wrote:Show nested quote +On April 22 2011 15:27 Hierarch wrote:On April 22 2011 15:20 Keilah wrote:On April 22 2011 15:14 Mykill wrote:On April 22 2011 15:05 Keilah wrote:
what math could I do? not hard calculations necessarily but you can always create custom game and time yourself using both strategies see where the advantages lie if any. this is just mostly guesswork/speculation. not very accurate i think. do you at least have replays of this getting you ahead? I sort of tried that, and to keep the data pure i built nothing but probes, pylons, and a nexus, but I realized that if I do that, i get enough minerals to make a new nexus much faster than i can reach 2 probes/patch, so nonstop production never has a chance to deviate from cutting at 2/patch. If I start building gateways and gas and units and whatnot, there's so much room for error where i just fuck up a little. I suppose that if I did some standard build and played near-flawlessly vs the very easy AI I could try both techniques and look at them in the replay. I'd have to be very careful to build the exact same buildings and units (other than a pylon or two), but it could be done. Just do a build, 2 gates, 2 gasses fully saturated and then constantly produce zealots, then one build you can cut probes and the other you not cut probes. OK so I did this. I did a standard economic build, no scouting, gate on 13, gas on 15, 2nd gas after building 19th probe, constant zlot production, build 2nd gate as soon as I have 150 minerals without cutting either probes or zlots, then pump probes and zealots with pylons nonstop and build the expansion nexus asap without cutting anything. I used all chronoboosts on probes, as soon as possible, and played pretty cleanly with few 'could have saved a half second here and there' moments in each game. One game I never stopped making probes. At 10:08 I had: -21 zealots (19 with 2 more building) -10 pylons -54 probes -96/100 food -1310 minerals -1632 gas The next game, I built probes until I had 17 on minerals, then cut until I got a nexus up, then resumed. This resulted in me storing surprisingly little chronoboost, I think I was on 35 energy or so when I restarted probes. At 10:08 i had: -21 zealots (19 with 2 more building) -11 pylons -60 probes -102/108 food -1560 minerals -1648 gas A net gain of 6 probes, 1 pylon, and 250 banked minerals, totaling 650 extra minerals by 10:08 for the game where I cut probes at 17 on minerals. Pretty nice! A few notes: -i built the geyser a tiny bit earlier in the 2nd game by accident, hence more gas at 10:08 -continuing to chronoboost probes after saturated seemed... retarded and unrealistic. It would probably be closer if I just stored the chrono for the probe-maker once he was saturated. -I didn't take a 3rd in either game, but if I had again stopped producing probes at 18 on minerals/base, I would have been able to build another nexus MUCH faster in the second game, presumably with the result that I would be even farther ahead if I'd expanded again. -In both games, I had 2 bases approximately saturated at 10:08 -I meant to stop at 10:00 but messed up a little, apparently if you save you can't load the replay so I was forced to manually count -because of my mistake I have no replays. But feel free to try it yourself!
It makes sense to cut probes to go fast nexus do to the chronoboost mechanic, however a key thing to note is if you cut probes you must chrono your nexus constantly to catch back up, however this can be not as good due to the fact that other things like to be chronod like putting 1 on warp gates or on a gateway or on hallucination etc... it's a tradeoff, but it's something to experiment with, due to chronoboost something like cutting probes which seems counter intuitive can work
|
Pretty sure probe saturation gives diminishing returns at more than 2 workers per mineral patch (ie 16 + 6 for the typical 8 mineral bases, 24 workers)
|
On April 22 2011 16:30 Severedevil wrote:
When you're saturated on one base, rally your nexus to your natural's minerals and distance mine.
w8 wtf?
|
ok, i just tested this in 2 games. one game i made a pylon, and cut probes at 16. i made a nexus at exactly 2:30.
second game, i built probes until 18, and i ended up having 400 minerals despite constant probe production, and got my nexus and second pylon down at exactly 3:00.
i dont know how you are wanting to do this, but i'm not really in the mood to test it anymore. although i'd say if you want to "cut" probes then you are probably thinking about making a nexus before 20 food amiright?
18 probes seems like its about the same as the 16 probe nexus, even though it's 30 seconds later, you dont ever have to cut probes, and you end up having a noticably stronger economy and you can make a forge, 3rd pylon and tons of cannons before you know it. by the time my forge finished i had a gateway, a gas, a 3rd pylon, a nexus, and enough money saved to drop 3 cannons instantly, almost 4. all of this with constant production of probeys
edit: and i forgot to mention, with the constant probe prodcuction i had exactly 24 probes when the nexus finished so i transferred over 8 and continued making probes. it actually seemed quite strong and given the situation would probably be ideal for a large economy based build
|
(Masters protoss). It seems to me to not be a question of economy, but rather one of your overall strength within the context of a game.
If you cut at 17, you're taking a riskier expansion. When it's complete, you'll have nearly no additional income, but you'll have a tactical and economic vulnerability sitting there. It's harder to defend than 1 base and you're now being forced to make probes off of 2 nexii with only 17 mining probes - a one base economy.
That leads to a 'juddery' rise in strength - you are not going to be able to add much infrastructure or military power while you're saturating the second base, and you're quite vulnerable at that time because you've lost your one base defenders' advantage.
Once it's saturated, then you're doing well, and it's economically more efficient.
Compare that to someone who was making extra probes on one base. They will take their expansion slower, which means they will have a greater army when they eventually go to secure it. Once it's finished, they will have an immediate rise in economy, both from the probe transfer and the fact they don't have to pump probes so hard to start to saturate. It's easier for them to improve the infrastructure needed and accelerate their military production appropriately.
There are also other advantages - if you get the opportunity to take a quick third, if you lose a nexus to a drop or bust of some kind - in both of these scenarios, it helps to haev the probes already on hand.
In short, it's smoother and safer to build extra probes and transfer, although it is marginally less economically efficient when compared to cutting probes for a faster nexus.
This is theorycrafting, however, and I could well be wrong in my assumptions. But I'd be shocked in the example above if the cut probes player had a larger army at any point from the 6:00-8:30ish period, and I'd be shocked if his army size during the most vulnerable period, taking the expansion wasn't significantly lower.
|
|
|
|
|
|
EPT
