Should Smogon have a comprehensive archive of notable, educational battles?

[XtrEEmMaShEEn3k2]

Edit: Or maybe some other site for that matter?

Being a bit of a chess enthusiast, I sometimes look through notable chess matches between famous chess players such as Kasparov, Fischer, Capablanca, Spassky, Alekhine, Karpov, and many others when I feel like delving into some strategies to improve my own game.

And I began wondering why Pokemon, with its deep strategy, did not have a comprehensive archive of notable battles where observers could really get a good look on how teams were used and played.

What I ideally envision is a tournament that was held in which all the participants would agree to have their battles archived and the details of the teams they were using (EVs, natures, and all) to be posted at the beginning of the battles so that future observers could get every detail on how the teams were set before the battle and how teams played out during the battle. Because more often than not, HOW a team is played can often lead to different results. Also, timestamps on those logs would be interesting to see how long a person thought about the move before making it, not to mention the actual HP Damage instead of just percents like it currently done in Shoddybattle logs.

Then, when the tournament is done, anyone could go on a site (Like say, Smogon, and include this in a new section "Battle archives"), and click on a particular tournament (let's say, "Infinity2000 Tournament X") and select any match in the tournament (Say, "Quarterfinals Obi vs. Jibaku") and really get a good understanding as to how the match played out and get a better idea on how to improve their own game, or just get a taste of some good Pokemoning.

I mainly got this idea from watching Chess videos by http://www.youtube.com/user/kingscrusher?ob=4 and http://www.youtube.com/user/Majnu2006, and imagining how cool it would be if someone could do that for Pokemon battles. You know, that maybe one day, someone would develop a program that would be able to reportray how a battle played out, instead of just posting battles between low-skilled battlers on youtube like what's going on now. And who knows, maybe someday someone will develop a program that would take in logs and reportray them like programs Kingscrusher and Majnu use in their videos to reportray notable chess games and other people would make videos about them.

We have warstories, but not everyone likes putting forth the effort of writing a warstory of battles they consider "Notable", and submitting a log to an online archive is much easier in comparison. Not to mention no warstory I've read has posted both players' teams' EV spreads, which I think would be a great help to observers who want a deeper understanding of the battle and both teams.

What do you guys think of this idea? And if you think it is a good idea, what do you think can be done to help facilitate it?

 

[Olie]

We do... it's called the warstory archive.
I'm not sure if a video log forum would happen, but it may soon with Platinum's battle recorder coming out.

 

[newbioform]

^Err, I don't think that is what he meant Olie. This is about a much more comprehensive and systematic approach than just archiving good warstories. Instead, we would mass-produce battle logs of different styles (with no commentary as those takes too much time) and file them under different categories to teach new players the way good battlers play and reflect the metagame shifts, something that really can't be achieved with one or two warstories archived per month.

That been said I cannot see how this would work out. Warstories are time-consuming to make and I don't think anyone would want to spend large periods of time looking through logs, which are rather hard to read and appreciate. Unlike the chess community where GMs and FMs actually analysis recent games through internet broadcast, the pokemon community is really just too small for this kind of stuff.

 

[Olie]

I'm still trying to comprehend what exactly this is describing. It seems to me like a request for an actual step by step thing that would take into account a list of both players moves and Pokemon, and see how it plays out, kind of like a warstory does. Unlike chess though where you can actually determine what will happen if you have a list of both players moves since there is nothing else included in the equation like dice rolling for example. (lol) Pokemon however includes Hax/items/random damage that make this nearly impossible since there are literally billions of possibilities.

 

[newbioform]

Bear with my jumpy logic:

An exact replicate of a battle shouldn't be a problem if shoddy can be programmed to record actual damage of each move. Similarly, it can be programmed to replay a match; this however would require significant effort from programmers for what seems like an insignificant sub-forum of the site.

Like I said, the pro battlers are not really likely to look through logs except for occasional reminiscence and new players who can actually understand logs (by which I mean getting the full picture with every move) are better off just using the time to play a few games on shoddy anyway. Warstories hold the advantage here as everyone can appreciate them effortlessly thanks to detailed commentary.

 

[gorm]

an archive of logs demonstrating a teams general/evolving strat is always going to be useful in one way or another (often explaining the effectiveness of a team) but it's not always worth managing for some people.

in a way team raters try to dream up situations where the team's strat would be hardest to execute and make fixes based on that.


the project you're talking about has pretty serious logistical work associated with it so don't expect it to be seriously considered for now but it's certainly interesting.

 

[Matthew]

So a team of: Infernape-Zapdos-Lucario-Salamence-Hippo-Skymin facing down a +2 Spe/Atk Gyarados and finding a way to come out on top by Intimidate switching and wearing it down with Sand Stream and to then finish it off by Extreeme Speeding by Lucario. Though to get Lucario in you need to have Zapdos or Salamence in so it bates a Stone Edge, and THEN you need to wonder how many Intimidate switches you have done and if your opponent has bothered remembering the pokemon you have switched to makeing it maybe a better choice to keep Salamence/Zapdos in to kill gyarados

like that?

 

[gorm]

...that's an example of a microstate of a certain battle, in particular one where you need to outplay your opponent to prevent utter destruction and wasn't really what i was talking about when i said "a team executing it's strategy". getting out of fixed with safe switches+resists is fine, but i was thinking more along the lines of

shuca heatran sets sr and explodes on a cbtar switchin
gyarados dds and baits a celebi switch, wears it down and dies
cm jirachi switches in and starts setting up, or switch to machap to catch tar/bliss

something like that, a series of logs which makes a sequence of events (or pattern of events) emerge. that sequence is pretty much you keeping the upper hand in a battle, and executing the teams strat based on your opponents weaknesses.

it would be interesting if we could develop statistics based on such an archive, eg, teams with zapdos tend to lose to teams with ddtar or something like that, which would cross reference+influence usage statistics.

interesting project idea but logistically insane considering the workload we have in front of us already x_x

 

[Wes]

I do like the idea, but it sounds like a lot of work:(

the pokemon community is really just too small for this kind of stuff.

I don't understand you there, I've been pretty amazed by how big the pokemon community is

 

[newbioform]

I don't understand you there, I've been pretty amazed by how big the pokemon community is

interesting project idea but logistically insane considering the workload we have in front of us already x_x

Well if we devote as much manpower into this as we put into C&C we will probably be able to organise a neat archive with many interesting battles recorded and formatted ready for statistical analysis, but is it worth while to do that? (badge for log filing lol) If we had that much manpower we are probably better off just setting up more tutors, which are a lot more effective than logs.

I do hope the decision makers consider this idea though, just when the pokemon community is bigger and everyone seems to have too much free time and nothing to contribute.

 

[XtrEEmMaShEEn3k2]

I personally would be willing to work on whatever it takes to get this going, but I don't know much about programing or doing anything with websites or whatever. But if someone told me what to do, I would probably do it.

 

[zarator]

I think it is pretty interesting, also because I love competitive chess and I completely understand what you mean:-) So I'll try to explain more deeply what - I hope - you want to create.

Lets say I want to create a new team. This team has a CB Ttar as a lead. Now, I ask to myself, "How does this lead specifically fares in the metagame". I pick said archive and tell to the browser: "Show me all the matches where one of the opponents lead with a CB Ttar". Now I have the opportunity to study all the possible matchups this pokemon could face. If I want something more specifical, like "How does CB Ttar fares against Hyppowdon?" I can simply ask the browser to narrow the results properly.

But this sort of method would not be limited only to leads. Say I want to study how Blissey matches up against Cresselia. I tell the browser to find all the battles where one of the opponent had a Blissey and the other one had a Cresselia. And there is a good probability that I can find many interesting outcomes to study and learn to how best play my pokemon.

This system offers the advantage to let people study not single pokemon, but, most importantly, the specifical situations that will occur into a battle with your team. It would complement very well the Strategydex and so it should be considered for every good battler

 

[TODICHUpwns]

sounds interesting but would take some work

also do we do it on our own logs or do we do it on someone elses log

 

[scorchedsky]

The idea of being able to cross-reference battles would absolutely be intriguing, but how would you narrow it? ie you pick the Cress/Bliss matchup. How would you get rid of every log that simply had Bliss/Cress switch in on each other and leave? The problem with picking the middle of battles is that they become too team-dependent. No longer is it Bliss/Cress, it's team v. team at that point

With lead v. lead, it'd be very interesting. Infernape leads would be able to see just how frequently Bronzong leads decide to stay in on them hoping to lay down SR, or something like that.

On second thought though, the searchable battles would be convenient even if it was just applied to warstory archives. We already have a lot of data (well maybe not yet for Platinum), so we might as well figure out a way to run a log through Warmachine or a similar program and get an output of which pokemon was out at each time. That database could be searchable, and then you'd read a log that was particularly applicable to whatever kind of matchup you were looking for.

 

[zarator]

The problem with the warstories is that they are written to entertain, not to educate the reader, most of the times.

Think to a commented chessmatch. First of all, the commentor is almost never neither of the battlers. His comment generally are dry and simply analyze the single moves made by the 2 battlers, telling why they did it, if they have been wrong or not, what should have been a better move telling a probable outcome etc, just like chess. Ofc, chess is much more predictable - a queen can make x different moves in a certain situation, cant randomally crit or freeze^^ - but it can be done to a lesser extent. Most of the matches, obviously, will not be commented, though, and are there simply to give people an insight of certain scenarios.

Commenting matches should be a lot more similar to RMTing, with the only exception that these comment must be made by only one person per comment - probably there would be a specifical badge for this, or the RMT badge would authorize also to make such comments.

 

[scorchedsky]

The idea with a chess match is inherently different, though, as every possibility can be analyzed at any time. You know which moves are possible from an outside observer standpoint, as well playing on either side. However, in pokemon, an objective observer either must be able to see both teams (if you support the RMT format), or neither (like it is in Shoddy). If you can see both teams, your judgment/comments will be different. For example, you wouldn't say launching random Fire Blasts is a good move if you know the opponent has Heatran lurking in the wings, but that's knowledge that only one side can possess.

 

[zarator]

This isnt a comment to a football match. Obviously the commentor will write about an ended match, knowing well which were the 2 teams. So, he will be able to say: " Player A here made a mistake. Vaporeon took less damage than Scizor from Ttar earthquake, but Scizor could have come in and force a switch with the threat of a Bullet Punch and set up Swords Dance". This is the kind of commentary im looking for

 

[XtrEEmMaShEEn3k2]

Well, I don't really intend for something as in-depth as cross-referencing possible matchups.

I mainly made this topic out of frustration of the fact that no logs of matches in past tournaments are really posted anywhere that is really easy to find. I'd mainly like, at the very least, for Smogon to post all of the battles in its tournaments.

 

[zarator]

Well, I don't really intend for something as in-depth as cross-referencing possible matchups.

I mainly made this topic out of frustration of the fact that no logs of matches in past tournaments are really posted anywhere that is really easy to find. I'd mainly like, at the very least, for Smogon to post all of the battles in its tournaments.

Well, maybe you did not intend it, but it would be very useful. And it could provide, maybe even better than anything did so far, that the way we take on pokemon here is not meant for certain noobs that lurk on the site^^

 

[newbioform]

Well, I don't really intend for something as in-depth as cross-referencing possible matchups.

I mainly made this topic out of frustration of the fact that no logs of matches in past tournaments are really posted anywhere that is really easy to find. I'd mainly like, at the very least, for Smogon to post all of the battles in its tournaments.

That is actually quite good. The tournament directors can just take a few extra minutes at the end of each round to dump all the logs they received into the archives. They are just txt files, an external drive would fit all the logs years into the future. While these logs make little difference right now and might not attract much viewing, they don't take too much effort either. This way when we actually decide to build a nice archive, we will have a large amount of data to work with at the start.

 

[MythTrainerInfinity]

In checkers some games are documented and new draws, wins, and losses are found as a result. Ex. http://www.jimloy.com/acf/bul-5.htm

For the lesser experienced or people who want to put a little extra study into something this would be a brilliant idea.

It takes time to get into pokémon competitively. There is a bit of a learning curve to understand the mechanics and seeing everything move for move would help getting into this easier for many. Some people learn by being explained a simple concept, others through study of raw move by move data. Either way I think this would greatly benefit Smogon as being the resource for information on competitive battling.

 

[eric the espeon]

I think it is pretty interesting, also because I love competitive chess and I completely understand what you mean:-) So I'll try to explain more deeply what - I hope - you want to create.

Lets say I want to create a new team. This team has a CB Ttar as a lead. Now, I ask to myself, "How does this lead specifically fares in the metagame". I pick said archive and tell to the browser: "Show me all the matches where one of the opponents lead with a CB Ttar". Now I have the opportunity to study all the possible matchups this pokemon could face. If I want something more specifical, like "How does CB Ttar fares against Hyppowdon?" I can simply ask the browser to narrow the results properly.

But this sort of method would not be limited only to leads. Say I want to study how Blissey matches up against Cresselia. I tell the browser to find all the battles where one of the opponent had a Blissey and the other one had a Cresselia. And there is a good probability that I can find many interesting outcomes to study and learn to how best play my pokemon.

This system offers the advantage to let people study not single pokemon, but, most importantly, the specifical situations that will occur into a battle with your team. It would complement very well the Strategydex and so it should be considered for every good battler

I think that this would be incredibly useful, but.. it would be a bit of a feet of programing to achieve.

Another thing to note is that whoever runs a shoddy server (Doug in this case) has access to every single battle log, that's how he makes the usage stats. SO we would not need to rely on people submitting logs. If those logs were made public and someone came up with a search mechanism (I have no idea how complicated this would be.) this idea would be possible.

Maybe if it was search able by having both players above a certain rating so you only see the battles with better moves made?
And to avoid the situation where you have to check ~10 logs where Bliss or Cress switched out it could be set up so that you could find a bit of log that ended with Bliss/Cress fainting and the other in battle.

 

[1234gjy]

Sounds like an amazing idea, but it's just too much work.

 

[Wes]

I still like the idea, but there's one thing to note: the metagame changes, alot.
I think suicide leads are a great example, because how many of the early teams were prepared for them. Or Heatran, most people said he was doomed with his "bad" typing. And I think there will be some more trends, trends we can't even imagine now. You may ask yourself:"hey why didn't that guy use thunderbolt on Gyarados?" which would now be pretty logic, but in the early days of DP, Electivire wrecked havoc (or tried to) and it would be stupid to do that. The main thing I'd like to say is that the time of the match may be important for such things.
I hope some more people would pick this whole recording idea up.

 

[Darkflagrance]

Even if the metagame changes, it tends to be cyclical. Some things are proven inferior, but others simply fall out of style, and then are 'rediscovered.' Just because a tactic is no longer used does not mean that it is no longer worth anything, and someone who reads an old warstory could conceivably come up with a new tactics based on older ones.