EPTOne obvious contender is information you were missing during the game.
What else?
Replay Review Process - What do you do?
| Forum Index > Brood War Strategy |
|
ajfirecracker
United States63 Posts
One obvious contender is information you were missing during the game. What else? | ||
|
WGT-Baal
France3531 Posts
I take it you mean a replay of yourself (and not like a pro). 2 branches depending on what i was going with: 1)all-in/timing: did i miss the timing? Why? Note key "structural" issues (eg you supply block yourself while scouting or something) and not so much micro failures. If i didnt miss the timing, and no obvious micro fail: what did they do and how could I delay that? 2) longer game plan or opponent allin: check macro, check scouting, should I have seen things coming, where to branch off my build? Things like that | ||
|
CicadaSC
United States1985 Posts
Other than that once your builds are tight you basically just look for the moment you lost and figure out why, what went wrong. A timing push killed you? Check where you were spending your money, your scouting, and what things you could trade off instead so you could have held it. that sort of thing, | ||
|
Navane
Netherlands2761 Posts
But also, if the opponent is a ladder Terran, when did he take his hidden third. | ||
|
art_of_turtle
United States1208 Posts
On August 23 2026 09:12 ajfirecracker wrote: When you pull up a replay to learn from it, what do you look for? One obvious contender is information you were missing during the game. What else? I started cheating and looking at Cwall to see what relative building timings are being hit. | ||
|
Slayer91
Ireland23357 Posts
The most important thing is that your decision making compared to an S rank player is guaranteed to be god-awful, replays allow you to watch those decisions in a slow paced environment with perfect information to try to unpack your terrible decisions so you can improve over time. Note: "Try to" is a very important qualifier here, as your brain that made the bad decisions is likely to miss the reason that most of them are bad, but some at least should be relatively obvious. The idea should be to breakdown your underlying assumptions and thinking that was incorrect so that the next game you will avoid or make less of the same bad decisions, this requires changes in your thinking which will be take time to digest so note taking may be a good idea as well so you can reinforce and review if you are still doing the wrong things again afterwards. If you are floating 2000 minerals at 10 minutes, you probably mis-managed your economy, you can always build extra production building or even two than theoretically optimal to make up for bad mechanics, but if you have that much extra money then you either are literally tunnel visioning a few units or you are not properly managing your macro. If you look a pro-gamer VODs, FPVODs or replays, check their supply compared to yours in quiet macro games, they are probably up by a base and 30 supply or something ridiculous by 10 minutes, for example if you see your own replays and you are sitting in your base getting lots of upgrades or going for weird attacks when you should be expanding and macroing up, that's what you need to be looking at and figuring out why you are doing it. Are you managing your economy badly? Are you afraid of the opponent having more than he should? Did you already fall behind and didn't realize how? etc I can look at any of my own games and find tons of awful decisions and I bet anyone can send me a CWAL replay of themselves and I can do the same thing. Any time you copy a pro build, rather than comparing the build timings - it's very likely that the biggest problems start immediately whenever you start making your own decisions after, like if you go ZvP 3 hatch spire into 5 hatch hydra, you can probably pull that part off without being too far behind where you should be, but if you stay on 29 drones and 3 bases and just build and micro hydras you will get stomped in mid game because you missed the part where you need to build drones and expand and at what timings. Maybe after the first replay analysis you see that, play another game, make the same mistake but realize it once you stop looking at your hydras and then panic and build 20 drones way too late and die to mid game timing attack again, then you need to look at that replay and again converge into the proper approximate timings for these decisions. (Expands, when to drone up, when to tech up, when to add on production etc) On the same note, imagine, you are a scrub, lose to protoss mid game army most games, you watch the replays, see your army get crushed, it's easy to think "WTF, psi storm is imba, I guess I need to build muta to snipe the templar or I lose" Then you tunnel vision into getting muta, have not enough other units, and still die to zealot/dragoon army after sniping templar from a protoss not even looking at them. That's because you are looking at artificial tactical things, you can easily compare your game to an S rank zerg and see that actually, he also loses units to psi storm, he just has 30 more supply so when he loses 20 units to storms he has 50 units left instead of 30 (this is HUGE, have you ever seen like 10 lings destroy 8 lings with 6 left over, imagine 50 vs 30 units, the swing that is in a fight), AND 10 more drones and an extra 2 hatcheries at an additional base so he can reproduce the lost army to recover after as well Sure, he will spend his larva better, micro better, multitask better, but your opponents will be so much worse it won't matter, if you can replicate a similar decision making you will win your games easily. Have you ever seen videos of like Artosis playing SnOW and just losing to mass scouts? That's an A rank terran, if you can lose to mass scouts it's bigger than a mechanical issue. (No flame to artosis, I've lost my fair share of games to mass scouts in previous years) | ||
|
Slayer91
Ireland23357 Posts
+ Show Spoiler [ASL Season22 | Ro.24 Group E spoilers] + Soulkey goes 12 hatch with a 2:52 lair (4 seconds later than optimal for larva timing as well as earlier mutas) against organ. There is a LOT of discussion here about how 11 hatch is better and 2:48 is the optimal timing you can hit with boosting. Soulkey probably is worse at boosting than Soma, and he also drone scouted at 11 which obviously loses you time but you are going to gain very valuable information. (Unclear if the 4 seconds is related to the drone scout purely, or boosting too, probably both) Anyway he crushed organ ( aka Ever)T(PianO if you are oldschool like me and recognize the name) without much fuss, it was a 2 rax academy bust which in theory someone could watch and go "oh having 4 seconds earlier Lair->Mutas here would be important to not die", but he just handled it calmly and played the rest of the ZvT perfectly as well. Rather than focusing on the timings, focus on what was important in this masterclass of ZvT -Having the speedlings scout to and confirm the build as medics/firebats confirm fast academy. -Having the threat of backstab speedlings to force firebat instead of an extra medic and gain seconds by distracting the terran -Pulling drones off GAS to build 2 sunkens and evo, realizing that minerals are the constraint in this situation, especially with 4th hatchery started -Evo chamber block to mess up the terrans micro, he EVEN CANCELS IT AND SAVES THE DRONE after (do not try this at home, just get the damn +1 carapace from it) -Pulling away drones from the second base to preserve workers, losing mining time is tiny compared to losing the drones and larva -Most probably target firing sunkens to focus fire two sunkens on one marine and the ones far from medics -Calmly microing 4 mutas against the remaining bio force (pulling away hurt muta) on an empty natural, since he can never kill the hatch with 5 marines in time -The overall game knowledge that you can simply allow yourself to pull drones off 2nd gas, get your whole natural busted and be ahead afterward if you play calmly, since he made sacrifices to get 2 rax academy out and lost his units ultimately, if he didn't actually kill your drones at your second because you pulled them, it didn't achieve that much. In comparison - when I get 2 rax academy busted I rethink my life choices, as it seems like you just lost early against a terran basically doing a normal macro build -The entire rest of the game has many many more nuggets of excellent decision making It's easy to go and miss all the important decisions HE made you would NOT here is the type of analysis people on TL.net might have made looking at a replay of someone getting 2 rax acad busted in the same spot [Greedy, high difficulty, iopq style] "Oh, delay the scout, work on your boosting, go for 2:48 lair, faster mutas, and you need 5 mutas instead of 4 to one hit marines" [Conservative, some scrub like me] "Oh make 3 sunks, even if you have to delay your third its worth not dying" It's also very easy to write-off the game on face value it as "Oh Koreans have god-tier mechanics, look at how he got his second base busted, barely holds with muta micro, and comes back and wins a macro game". That is not at all what actually happened. In our games, we lose all our 2nd base drones, don't have a 3rd base started, and play with 2 hatch 2 base 12 drones from that position vs 4 hatch 3 base with 22 drones (2 base 1 drone per patch). What actually happened is Soulkey is smarter than you AND has better mechanics, but you can steal all his hard thought conclusions and tactics from this game for FREE (minus ad watching time) Now, you won't save your evo chamber drone by cancelling and running it, and save the 5th muta with perfect micro, and your speedlings might die trying to scout, but almost everything else that he did, you can do too. | ||
|
iopq
United States1157 Posts
On August 28 2026 21:32 Slayer91 wrote: I strongly disagree with people here focusing on checking timings vs pro builds, or how much money you are floating, you should mainly be looking at decision-making, mechanics takes practice, you won't be able to correct bad multitasking by watching a replay. Seconds in build orders are not going to be winning or losing games for you at foreigner level. I just saw Stryker lose to a bunker rush because he put his pool down 2 seconds later than he could have and the marines had just enough time to enter it. If he had the pool just 1-2 seconds earlier he would have blocked it | ||
|
Slayer91
Ireland23357 Posts
I would also be curious if there was nothing else Stryker could have done to fight back the bunker rush other than having an earlier pool. Overall small differences in maps, barracks position, spawn locations, will cause too much variance in the bunker rush timings in terms of 1-2-3-4 seconds that I think focusing on late timings it's probably not a good way to improve. Lots of 12 hatch vs bunker rushes pro games involve drones delaying but killing some marines, trying to minimize drone losses by pulling back damaged ones, using good micro, blocking, drilling etc, and then very often the bunker finishes at some point and you have to time your ling+drone attack and micro against repairing SCVs to kill the bunker There are also decisions like, well what if terran does depot rax but bunker rushes as a fake to draw 8 drones, what's your strategy on that? Do you just keep 1 drone on the scouting SCV to prevent it finishing a bunker without marine support, and only pulling more drones when you are sure its 8 rax? What's 2nd overlord scout strategy? Should you leave it in a strategic location to scout marines and always send one drone when you get SCV scouted, or keep it close to natural to scout for bunker rush? (letting a bunker finish without your vision before hatch completes is obviously bad) I guess my fundamental point is that there are a variety of aspects that can be improved upon even in a short game (a lot more in a longer game), that writing off a loss as purely due to a late pool is actively detrimental to your improvement I go into every game trying and generally fucking up my boosting and getting everything 2-3 seconds late but it probably causes approximately 0% of my losses, even if it influenced 1 or 2 it was probably one of a very long list of things I did badly | ||
|
iopq
United States1157 Posts
On September 01 2026 19:05 Slayer91 wrote: I would argue that for the majority of players learning how to use drones properly is more important than your exact pool timing, you can't block bunker rushes with lings on a hatch first, obviously it helps if they are out faster, but they are always too late for 8 rax rushes, in that terran usually always can get the bunker up I would also be curious if there was nothing else Stryker could have done to fight back the bunker rush other than having an earlier pool he was using the drones, but you can't micro perfectly and block the marines completely off forever I looked at a different replay of his (I wasn't obsing the other bunker game) and he could get lings out exactly 7 seconds earlier by doing a slightly tighter build and boosting a little more efficiently (5 seconds from build and 2 seconds with better boosts). The downside is you can fuck it up and slow down your lair/float 3 larva for a second or two. FWIW, Soma does this 11/10/10 build almost every game vs. Terran and just boosts to hit the correct lair timing (only like a second or two slower than perfect most of the time) there has to be a reason he does that instead of the more economical 11/11/10 that gets another drone out earlier | ||
|
Slayer91
Ireland23357 Posts
But Soulkey did 12/11/10 in an ASL game with a normal scout timing, so obviously it's still playable at the top level and he's not losing every 8 rax game because his lings are 5s late. There's obviously trade-offs to everything, arguing about which build is optimal vs what is exactly the discussion I think is over-emphasised not because it doesn't matter but because there are almost always way more important factors that you can improve on. he was using the drones, but you can't micro perfectly and block the marines completely off forever Yes exactly, a lot of the time a marine or two marines will get into a bunker and managing that situation with a ling/drone bust on it is also an important skill to practice. There WILL be bunker rushes that get marines in before lings pop and 2 rax academy busts that bust before 2:48 lair mutas arrive, I'm suggesting focusing on how to handle those situations is more important that trying to figure out if you can sidestep the problem with slightly faster build, and I'm also saying that these situations are already a small % of games where the couple of seconds actually mattered, overall gameplay decision making is what decides much more games | ||
|
iopq
United States1157 Posts
there has to be a reason this is the new standard, and I suspect the reason is that all three of them do straight up lose to bunker rushes sometimes (in fact, every Zerg does) this is despite lair being 4s late(2:52), he still prefers doing it anyway | ||
|
Slayer91
Ireland23357 Posts
If terran is 8 raxing a lot the real solution is to overpool, the value of faster scout vs faster pool and hatch is something I have no idea about, there are probably ways the faster pool is more robust and other ways the faster scout is more robust. In random games there tends to be a lot less cheese as pros are practicing standard play and in tournaments the value of scouting goes up as almost everyone is preparing cheese/all-in builds. This is derailing the thread though, this is exactly the type of discussion I think is largely less important for improvement, but people love to discuss endlessly Whatever advantage you get on a 4 second faster lair or whatever in most games is much less important than if you can lose less mutas by selecting which packs of turrets and bio to take swipes on and which ones not to, for example, if you lose even one extra muta it becomes Of course you can cherry pick games where build timing matters, it obviously matters to some extent, I'm just saying its a very superficial thing to look at, just pick a build and get it pretty good, you will always lose seconds in real games anyway with any kind of harass or any wrench thrown in by the opponent, almost all games are won and lost on the back of mechanics and decision making (and again, mechanics is not really something that can be improved by looking at the replay) | ||
|
iopq
United States1157 Posts
and yes, you can improve every other aspect of the game, I've been playing this game non-stop for several years and I only improved from B to A since 2023 I disagree mechanics is even a big deal, I get smashed by 120 APM protosses who can hold S the exact timings are more important than being able to send drones to work more efficiently | ||
|
Slayer91
Ireland23357 Posts
The most important thing to check in your own replays is to do your best at analyzing decisions and deciding if they were bad and if so, why you made them. Essentially, your replay hides all the evidence of your bad play for you to uncover without being distracted by the flurry of a live game. What I believe is less important is: -Checking your build order timings, as I believe optimisations are insignificant when compared to glaring mistakes or poor skill level -Checking how good or bad your mechanics was, as in my opinion this is mainly about practice. This does not mean floating minerals = bad mechanics, it can also be mis-planning your economy e.g. worker vs hatchery count What actually disappoints me here is people acting like they are just checking one or two small things like how fast their timings were or whatever and going next, how will you ever improve that way? You are probably making 100 bad choices and you are not even trying to find out what they were Surely in many games you are encountering things that feel like "problems" and you want to look for things that are "solutions", and refer to pro games if you run out of ideas? There is an interesting bit about mechanics, which is do you need to train yourself to macro while doing other things, like microing mutas, microing a large fight, etc, I am leaning toward yes, but it will also take a lot of practice I kind of feel like a certain mindset and urgency is also required, like "I want to micro my mutas but I NEED to do this" | ||
|
WGT-Baal
France3531 Posts
When i was talking about a timing attack/allin though for me i didnt mean 3-4s but more 30s or insane deviation. It was geared more for lower ranks. Say you try a 3 base bust PvT but you max out in 13 or 14mins instead of 11:30, that s a problem. If nothing else happens that game you should work on that first. Now you are absolutely right that being ableto judge whether to even attempt that based on what T does is a key skill but i d argue it comes a bit later. And at pro level judgement trumps the rest (except perhaps snow s reaver micro) because they all have great mechanics. But for us down in B and below it s probably equally important. Now if you do not go for some sort of timing/allin then ues scouting and decisions are key. Ut s also a lot harder to practice specifically | ||
|
Slayer91
Ireland23357 Posts
Unless you are floating 3 mins worth of minerals you probably took late expand timings etc Actually, where I would look at replays for timings, is more "is this time a good/bad time to attack" as you can see exactly what the opponent has when you are merely guessing normally speaking When I'm talking about decision making, I'm not talking about deciding to do a build or not, obviously that is part of it But actually navigating the game to get to a maxed army involves tons of decisions like 2nd nexus timing, 3rd nexus timing, how to position your army to block harass, when to add gates/how many, how and when to add harass builds (like shuttle reaver) to buy time.. theres probably 100 decisions you make automatic or on a whim and maybe many of them are not good on close inspection. Like deciding to build a shuttle and reaver around the time the terran is already ready to push out, where its too late to harass, OK technically its a timing but thats literally minutes where you decided not to put down a robo | ||
|
WGT-Baal
France3531 Posts
You could consider all of that decision making of course, but if you ve decided to all-in "no matter what" and the opponent doesnt disrupt you step 1 is to at least nail the timing on your end. Once you can do that reliably then everything you say is key | ||
|
Slayer91
Ireland23357 Posts
Ok the replay can be used to confirm something like this but I would say its not the main purpose of reviewing replays. Single player practice is probably underused as well. | ||
|
WGT-Baal
France3531 Posts
On September 04 2026 07:30 Slayer91 wrote: Fair enough, but then, that's really a single player practice exercise, not a replay review activity? Ok the replay can be used to confirm something like this but I would say its not the main purpose of reviewing replays. Single player practice is probably underused as well. I d argue it depends. I do a lot of single player practice but online in a "real" game you have a bit more variations (could be stress, the TR, frustration with the previous game,the map etc). And if the opponent doesnt attack you it s still worth it to check what happened. Maybe vs a human you try to be fancier than vs the AI so you might overmicro your scouting worker, or first units, or overract to something minor etc. But again i mean those as step 1 (or 0). And also some matchups (PvT) are more like single player than others (zvz for one). You should review everything but to not get overwhelmed checking your mechanics first is good and always true, then the next steps are game and MU dependent. Another example: for me personally, I have mechanical/multitasking failures often in PvP 1gate core vs 2 gate. When the zealots arrive i often either miss probes, miss a pylon, overmicro, delayed 2nd gate or range etc... those are all mechanical issues you cant really train single player. Another question for you though to keep the thread moving: what do you do when you watch a replay and genuinely do not know why the opponent is so ahead at a given time? try to find similar openings on the same map in a pro VOD? | ||
|
Slayer91
Ireland23357 Posts
On September 04 2026 22:30 WGT-Baal wrote: Another question for you though to keep the thread moving: what do you do when you watch a replay and genuinely do not know why the opponent is so ahead at a given time? try to find similar openings on the same map in a pro VOD? Yeah, when I can't figure it out I tend to binge watch pro videos and find similar games, I feel super stumped quite often. Ultimately I'm noticing that the devil is generally in the details, so if you see the pro guy doing a similar thing overall to what you're doing, you have to deep dive into the replay to notice things you are doing wrong, usually bad economic management, in my experience. In fact, I rarely see something crazy different to what I was trying, or else there are several options and they all work if well executed. It's kind of crazy when the game seems so mapped out how often I lose to something and feel stumped, but often those are the best things to analyze as it's a way to find proper gaps in your understanding of the game. I had an issue with goliath mech terrans in ZvT, I just had to learn to be way more greedy as they have no early pressure, but also not then over-greed and prepare for the timing push. What was happening is I was over-defending like against a fictional bio terran, and then over-defending a ~9 goliath push which can kill you sometimes, and then keep making units as well from there and droning way too late You just need to barely defend the low number of goliaths (vision is important, burrow/overlords/lings etc) and mostly greed to defend the stronger 10-11 minute push with 24 goliaths and 3 tanks. But I saw in CWAL replays many other Bish rank zergs, while playing similarly, manage to beat the same builds playing a bit more aggressively using mutas to trade their hp for SCV kills (goliaths and turrets kill mutas slowly without focus fire) and using more sunkens to delay, for example, so several styles work (muta/hydra with mostly hydra vs mass muta and sunken first) | ||
| ||
