Microlimits needs a library

[Aaron W.]

And some librarians... This is the sort of project that would really benefit the entire forum. But I need some help getting it started. I don't know if this project will go anywhere, but it's worth a shot.

Overall picture - The search function works, but it's a little hard to get what you really want out of it, and many times you've got to wade through far too many posts to find things. I would like to see something like a library, where you can go in and click around to get to certain shelves which contain the posts which are actually relevant to your question. This would be nice for quick reference and reviewing common situations.

But there are a few steps between here and there.

Step 1 - Library Orgainization

There are so many ways of sorting hands:
- Hand type (AK, JJ, T9s...) and concept (Equity, free card, slowplaying)
- Position (Early, Middle, Late, Blinds)
- Street-based decision (Preflop decisions, flop decisions...)
- Situation (Overcards, PP flops overcards, Flush Draw, Weak Draws, Heads up...)
- Lines (call-call-call, call-call-bet, bet-fold, stop and go...)
- Others (Ed Miller essays)

This is hard because I don't know how best to orgainize these things into reasonable categories. It could be something like a Dewey Decimal System, but that's cumbersome because of all the overlap of concepts. My idea is that hands would be flagged with certain properties which could somehow be searched and sorted, which brings us to...

Step 2 - Library hosting and access

Someone would need to put together an accessible database (someone who is fluent in web programming would be nice -- I am not such a someone!). I imagine being able to "browse" the library shelves and to search for hands which meet the flagged criteria from above (KK + overcard on the flop, Axs + flush draw, 22 + set). Once the organizational aspect is taken care of, I think it will be much more clear what to do here.

I imagine only getting a series of links which point back to the 2+2 forum server (okay, it's more like a card catalog than a library), . It would be silly to store the entire posts when someone else is already doing that.

Step 3 - Library maintainance

Once it's up and running, it would be good to have a team of people who will keep adding entries into the library (how many? I don't know.). It would be on the programmer in step 2 to make "librarian" access available to those who want to do it. They would need to figure out how to do it in a way that doesn't end up with duplicate entries.

I guess one person could do this alone, but that would be an everyday task that most people won't want to commit to.

Here's what is needed *NOW* in order for this project to get off the ground: What is the best way of organizing posts? How can we make it so that it will be easy to find something you're looking for? Just think about the types of things you want to know and the sorts of parameters you would want to use to get to that information.

[Isura]

This is a very good idea. To get this started though, my idea is to start a new thread where everyone can post the links to their favorite posts (eg. classic Ed Miller posts, etc to get it started). From this, hopefully a natural structure to organize the posts will arise. I'm no expert in web hosting, but I assume that since most of the contents will be URL's to 2+2 posts, space shouldn't be an issue. I'll post some of my favorite posts here to get it started.

[Isura]

Here's some useful stuff I dug up. The categories in the parenthesis are just guesses for organization purposes.

You play too tight – Nate (general theory)
http://forumserver.twoplustwo.com/showflat.php?Cat=&Number=1767198&page=&view=&sb=5& o=&fpart=1&vc=1

5 things LL players should "unlearn" – Clarkmeister (general)
http://archiveserver.twoplustwo.com/showflat.php?Cat=&Number=321233&page=&view=&sb=5&o =&fpart=all&vc=1

Charging flush draws – Ed Miller (general theory)
http://forumserver.twoplustwo.com/showthreaded.php?Cat=&Board=micro&Number=459928&Fo rum=,,c2,,&Words=&Searchpage=2&Limit=25&Main=45930 1&Search=true&where=bodysub&Name=&daterange=1&newe rval=2&newertype=y&

QQ on K22 board – Ed Miller (standard line)
http://archiveserver.twoplustwo.com/showflat.php?Cat=&Board=&Number=969383&page=&view= &sb=5&o=&fpart=all

Preflop Play: Take off the training wheels – Ed Miller (general theory)
http://archiveserver.twoplustwo.com/showflat.php?Cat=&Number=495010&page=&view=&sb=5&o =&fpart=all&vc=1

How to make river decisions – GuyOnTilt (theory and hand example)
http://archiveserver.twoplustwo.com/showflat.php?Cat=&Number=560235&page=&view=&sb=5&o =&fpart=all&vc=1

[mhardy]

A great idea Aaron.

I'd like to get my name on this thread to volunteer for whatever (I don't have a web site/html skills but am willing to do whatever grunt work someone might come up with). Anyone email or PM me here.

Maybe #1 could be categorized under 2-4 skill levels. A "beginner's" top level wouldn't need to be concerned with equity for instance. Maybe that kind of categorization would lend itself to sorting the concepts? Just a couple branches at the top spreading out from there...

[MrWookie47]

I think the best way of organizing it would be with two designations: one for hand type, and one for opponent type. That's not too extensive, but it would cover many basic questions. Hand types could be (not necessarily limited to) overpairs, top pair, second pair, weaker pairs, sets, trips, flush draws, straight draws, unpaired overcards, flushes, straights, and monsters, as well as including "general theory," and "preflop." Opponent types would likely be typical, unknown, loose passive, LAGs, TAGs, rocks, and maniacs, or perhaps just use all of the PT definitions (sLAP, sLAA, TPA, etc.).

That would cover the most basic queries. I might also be tempted to add a third and fourth designation: two booleans, one for the board, simply coordinated or uncoordinated, and one for heads up or multiway. I'm not sure if that would be useful, or going overboard. That's sort of why I didn't have any additional elucidation of the preflop category. Since preflop play is as black and white as it gets in poker, devoting so much search to it is kind of pointless.

A search feature for this library could then have a set of check boxes for what hand types you want to search for, and with what opponent types (possibly two sets of radio buttons for the two booleans, plus an "unspecified" option). A user could check the boxes of interest and get back a series of pages that match any or all of the criteria specified.

I realize I'm neglecting position-related searches, since that is important. However, I would hope that eventually, in a populated library, a query for a hand type and opponent type would yield results of plays made from many positions, giving the reader an opportunity to see how play chages for that situation based on the position.

A more complicated system may let players find a hand post that very nearly matches his or her own situation, but I'm not sure that would be better. There's always the KISS philosophy to consider, plus the fact that looking at many related, yet not identical, examples may help you to better consider the various factors going into your decision, rather than just finding out how exactly to play a specific situation should it ever come up again. This is also why I think specifying play by a particular street is too much. So much of what you do on a given street depends on hand type, opponent type, and what you did on the previous street, that sorting by "Turn issues," I think isn't as useful. Also, SSH does a fine job of covering general issues relating to general play on each street.

The more I think about this, the more I wish it was implemented in the 2+2 search feature already. Good idea.

[MrWookie47]

Just to add on to my previous point, I made that post with the mindset that this library isn't there to spoonfeed players all the right answers in every situation. Instead, it would be set up with a goal to give players the tools to find many different situations to consider, read their consensus answers, and then to synthesize the considerations into their own personal question. Or, if they can't find an acceptable set of related situations, they then post the hand for discussion. If this library's goal is to have more of a spoonfeeding role, then a more specific search function will be necessary.

While I'm thinking about it, keeping a database set up in the above manner wouldn't be too ridiculous. It would require a dedicated, web-capable person or persons (sorry, web-capable means not me) to categorize the days' posts, and add the link to the database, along with the relevant category information and the post title. It's not easy, given the volume of posts, but keeping the number of variables small will sure help out the people responsible.

[btspider]

i suggested this the last time it came up.. this would be if no external system would be introduced and we had to work with the forum structure here.

create a master favorites thread that has links to all the categorical favorites thread.

each categorical favorite thread would have a specific focus (that could further be broken down). anytime there is a favorite worthy hand/topic, it'd be linked to in the appropriate subtopic.. but we'd only need the reference to the master topic to find anything.

e.g.

Topic 1: Master thread (has links to topic A, B, C, etc..)

Topic A: overcards (links to favorite worthy topics + links to topic A1, A2, ...)

Topic A1: PFR'd overcards that should not be auto-bet on the flop
Topic A2: overcards that are UI on the turn..

Topic B: big draws

Topic C: HU lines..

etc..

[KaiShin]

Could such a thread be flagged so it could be perpetually editable? Cause it would suck to continuously have to post updates and stuff instead of just changing the OP.

[Aaron W.]

It's a good idea, but I worry that it may get difficult to use as the chain gets longer. I see that turning into a strange maze of links that's hard to move through as the chain gets longer and the number of sub-classes increases. It will also be hard to work with updates when new favorite posts start popping up. This also has a management problem when people start putting in things that don't really belong, and so forth... (but I leave it open for you can convince me that it won't happen that way)

This still has the same classification problem that my system has. What are the categories of things that we want? How should it be organized? I think that's the bigger hurdle than the nuts and bolts of how we put it together.

[Stuey]

My head hurts thinking about trying this.
It is a great idea but lets look for the easy answer.
I don't know much about poker or programming.
But I know how to get out of doing hard work!!
Here is my idea I don't think it is the answer but
it might push the great minds here in the right direction.

The search function does not work well enough.
There is a ton of information being rehashed because the
Original discussions get lost.
So how can we make the search function work better?
Build a new one? NO too much work
Change the way we post so this one works better? Yup yup yup

The subject line of a post is important! Don't waste it
with stuff you think will make you sound cool.
Let's standardize it! Here is an example

(Table position of hero) (raises/unraised preflop)
(hero's hand) // (hand identifier) I need poker expert to
detail these.

MP1 Raises KK // Turn paired board HU

Anyway the idea is you cannot make up your own subject line.
You have to pick from a glossary of terms provided that suit
your example.

This would become our dewy decimal system!
It needs to be setup by someone who knows poker well.
It needs to be used by the posters.
It will encompass a huge number of poker situations but
that is ok we are going to build this library one hand at a time.

Next we have to identify good threads. We need a watchman.
Threads that have every idea about the situation
explained and examined. We need to actively post in these
threads to raise the quality of them not me but smart guys.

When the librarian thinks a thread is good enough to save he
puts a link to it in the main thread.
They are listed in order like this or close to it.
Depends on our system.

MP1 raises KK // turn paired board HU is before
MP2 raises KK // turn paired board HU which is far above
MP1 raises 99 // flop no trips muti way
This is our card file!

When some joker posts a hand that has been covered
We flame him and send him to the saved thread
If someone does not post the hand using the system
We do not reply to it.

I think I stole btspiders idea!
Anyway I like the idea of the library.
But I’m to lazy to help much.
Make a new thread for this the links above are killing me.

[btspider]

[ QUOTE ]
It's a good idea, but I worry that it may get difficult to use as the chain gets longer. I see that turning into a strange maze of links that's hard to move through as the chain gets longer and the number of sub-classes increases. It will also be hard to work with updates when new favorite posts start popping up. This also has a management problem when people start putting in things that don't really belong, and so forth... (but I leave it open for you can convince me that it won't happen that way)

This still has the same classification problem that my system has. What are the categories of things that we want? How should it be organized? I think that's the bigger hurdle than the nuts and bolts of how we put it together.

[/ QUOTE ]

i agree its a potential logistical nightmare and my programming mind thinks its terrible. but sometimes something is better than nothing. we'd have to use an outside resource to come up with a better solution.. e.g. a wiki or something which links back to 2+2.. with a few designated managers.

[Scratch]

Hey guys-- I've been lurking here for a while and have thought about having some sort of searchable database of hand/theory discussions. I am also a programmer and I think one approach would be to somehow hook into bison's or another converter and embed metadata. The metadata would describe things about the hand (type, position, and even some generic situations). A lot of this metadata could be generated automatically, as part of the conversion process (although things like equity, free cards, etc. might need to be manually designated by the poster).

The next step would be to create a hand search page where the user selects the criteria they're looking for. It's then a simple matter of searching the forum for the appropriate metadata. Simple in theory, at least. We'd likely need an external website to host the search page.

[Aaron W.]

[ QUOTE ]
Hey guys-- I've been lurking here for a while and have thought about having some sort of searchable database of hand/theory discussions. I am also a programmer and I think one approach would be to somehow hook into bison's or another converter and embed metadata. The metadata would describe things about the hand (type, position, and even some generic situations). A lot of this metadata could be generated automatically, as part of the conversion process (although things like equity, free cards, etc. might need to be manually designated by the poster).

The next step would be to create a hand search page where the user selects the criteria they're looking for. It's then a simple matter of searching the forum for the appropriate metadata. Simple in theory, at least. We'd likely need an external website to host the search page.

[/ QUOTE ]

Hooking it into the converter may be more difficult than it's worth. There are also lots of posts that don't go through the converter at all (in fact, many of the better posts are general theory discussions that don't even have a hand in it).

I guessed that this whole deal would need be housed offsite and that the search engine is not tied to the 2+2 servers (I don't know what that sort of access entails). I now regret labeling it a "library" because it's definitely turning into a "card catalog".

The way I see it working is that we can have a team of two or three people who can manually enter posts into the library. I imagine a simple web form with fields or pulldown menus for

- Subject
- URL to the original post
- Type (Hand discussion/general theory/other)
- Hand type (if applicable)
- Position (EP/MP/LP/Blinds)

For the other categories, it could be a series of boxes to check if it has that property:

- Free card
- Equity
- Slowplaying (Don't do it!!! /images/graemlins/grin.gif)
- Waiting until the turn
- Buying outs
- Out of position
- Playing overcards
- Playing a pocket pair facing overcards
- PT Stats
- And so on...

Is something like this workable?

[tiltaholic]

Aaron-

You know I've been thinking about this too.

I have a lot of comments but here is the condensed version, stream of consciousness style -

- this is a huge, huge, epic undertaking.

- this is a job for someone(s). an almost full time job. i mean, i don't read all the threads in micros anymore. there are just too many. and the good ones take me hours to read, days to think about and decide that they are good, and weeks to realize exactly how good and useful they actually are. then again, i basically suck at poker.

- what makes this board great is the -conversations- that arise, as situations arise. not just the written info. don't get me wrong, i know there is a TON of info in the archives, and tons of good threads that are posted here every week. but the interactive nature of the boards is one of the primary reasons for their collective success.

- it seems to me that part of the "organizational scheme" planning is a desire on the part of some of us to control something that is, by its very nature, uncontrolable.

- now there is a teacher in me. i teach. i like to teach. i like to help people understand things, as far as i understand them myself. and i want to learn more myself too. but this project is monstrous. and you know what, i also want to improve my game, and, god forbid, maybe someday move out of microlimits.

- and the selfish in me says this : you know what - success in poker (or anything else) doesn't come care packaged and tagged for their easy use. you need to work to learn. it takes time and effort. not the tireless work of 3-10 dedicated micro cataloguers and system engineers. at the end of the day - this place is just what it is. a great resource of information - that you can mine for yourself.

- finally as i mentioned before, i will be happy to help - and hopefully will be collecting some good threads of recentness to post.

[Scratch]

You're probably right-- messing with parsers can be a pain. Also, I don't think every thread should have the same weight. Some discussions are obviously better than others and should be prioritized as such.

As for your suggestion, I don't think that it would be terribly difficult to implement. The hard part would be, as you say, in defining the searchable categories. Also, what happens to old threads that are archived? I assume that their URL changes? Just something to think about.

The technical aspect shouldn't be too bad, but the administrative side could be /images/graemlins/ooo.gif.

[JayD]

Maybe I'm going in a different direction here, but here is my idea.

I think a wiki functioning as a community bookmarks system for 2+2 would be perfect for what everyone seems to be shooting for. It could link to informative threads on 2+2 so that they aren't buried. This site has so much content and the search function is so inadequate it needs something like this.

The wiki could also function as a general poker knowledge base, including pages on basic concepts. Whenever someone asks a simple question you could just point them towards the wiki. Of course, the page couldn't contain any copyrighted material. Here is an example of a knowledge base type wiki which anyone can edit: Emacs Wiki (http://www.emacswiki.org/cgi-bin/wiki)

Regarding managing the wiki: why not allow anyone to edit the wiki? Large wiki systems like wikipedia and emacswiki function just fine with the "anyone can edit" policy. There is already a huge 2+2 community, why not take advantage of that?

[btspider]

[ QUOTE ]
I imagine a simple web form with fields or pulldown menus for

[/ QUOTE ]

this check-box form could be chucked into the converter or a separate website and just create the meta-data that would be cut/pasted back to 2+2.

i think once you are dealing with an external search engine/data, you run into a lot of issues (what happens when the posts move to the archives and the links are not updated, etc)

the meta-data would be pretty straightforward, something like:

in white at the bottom of the post:
UTG/EP/PP/JJ/HPFR/OCFLOP

this would be searchable here simply using a:
-re: +PP +EP +HPFR +OCFLOP

(list of Hero EP PP hands that Hero raised PF and faced an overcard on the flop)

[Aaron W.]

[ QUOTE ]

- this is a huge, huge, epic undertaking.

- this is a job for someone(s). an almost full time job. i mean, i don't read all the threads in micros anymore. there are just too many. and the good ones take me hours to read, days to think about and decide that they are good, and weeks to realize exactly how good and useful they actually are. then again, i basically suck at poker.


[/ QUOTE ]

It's not necessary for *ALL* posts to be entered in the catalog. Even if 5-10 posts get entered in a day, there will still be a couple hundred by the end of the month.

[ QUOTE ]

- what makes this board great is the -conversations- that arise, as situations arise. not just the written info. don't get me wrong, i know there is a TON of info in the archives, and tons of good threads that are posted here every week. but the interactive nature of the boards is one of the primary reasons for their collective success.

- now there is a teacher in me. i teach. i like to teach. i like to help people understand things, as far as i understand them myself. and i want to learn more myself too. but this project is monstrous. and you know what, i also want to improve my game, and, god forbid, maybe someday move out of microlimits.


[/ QUOTE ]

I don't think this will kill the conversations. I view it as a tool to help teach and review. By having posts organized in a clear way, it's easier to help walk people through the learning curve of becoming a better poker player. If someone is really struggling on a certain point, or doubts the merits of a certain play, this will help pull out examples of past hands and discussions that illustrate the idea.

[ QUOTE ]

- it seems to me that part of the "organizational scheme" planning is a desire on the part of some of us to control something that is, by its very nature, uncontrolable.


[/ QUOTE ]

That's why I see this being an external page, so that (this part) is contollable. I do know what you mean, and I have the same sort of frustration with this board as the new thread count continues to climb, but I feel that there must be something that can be done to reduce the clutter.

[ QUOTE ]

- and the selfish in me says this : you know what - success in poker (or anything else) doesn't come care packaged and tagged for their easy use. you need to work to learn. it takes time and effort. not the tireless work of 3-10 dedicated micro cataloguers and system engineers. at the end of the day - this place is just what it is. a great resource of information - that you can mine for yourself.

[/ QUOTE ]

Well, this might be more of a philosophical point. If a student is lazy, making the material more accessible will only make him moderately better. It still takes a good deal of work to learn to think poker, and skimming over a bunch of posts on the same topic will not make that happen. I doubt that if you just read through all of the KK with an ace on the flop hands that you'll play it any better unless you understand the *WHY* of each variant. It's just like people reading SSH once and saying "I've read that" as if that's the end of it.

[ QUOTE ]

- finally as i mentioned before, i will be happy to help - and hopefully will be collecting some good threads of recentness to post.

[/ QUOTE ]

Hopefully this is a goal that's within reach... I do recognize the size of it.

[Aaron W.]

[ QUOTE ]
i think once you are dealing with an external search engine/data, you run into a lot of issues (what happens when the posts move to the archives and the links are not updated, etc)

[/ QUOTE ]

As far as the links changing, is that an issue here? What happens to posts that include a link to another post as a URL? I can't imagine that those links get changed, yet I've never run across a post where the link failed to go where it should go...

[btspider]

[ QUOTE ]
[ QUOTE ]
i think once you are dealing with an external search engine/data, you run into a lot of issues (what happens when the posts move to the archives and the links are not updated, etc)

[/ QUOTE ]

As far as the links changing, is that an issue here? What happens to posts that include a link to another post as a URL? I can't imagine that those links get changed, yet I've never run across a post where the link failed to go where it should go...

[/ QUOTE ]

ah cool.. i was wrong. found an old post of mine with a link to an archived topic (which wasn't archived when i linked to it).. the link went to forumserver.twoplustwo.com and got redirected on the backend to the archive server. nevermind.

[Aaron W.]

[ QUOTE ]
Maybe I'm going in a different direction here, but here is my idea.

I think a wiki functioning as a community bookmarks system for 2+2 would be perfect for what everyone seems to be shooting for. It could link to informative threads on 2+2 so that they aren't buried. This site has so much content and the search function is so inadequate it needs something like this.

The wiki could also function as a general poker knowledge base, including pages on basic concepts. Whenever someone asks a simple question you could just point them towards the wiki. Of course, the page couldn't contain any copyrighted material. Here is an example of a knowledge base type wiki which anyone can edit: Emacs Wiki (http://www.emacswiki.org/cgi-bin/wiki)

Regarding managing the wiki: why not allow anyone to edit the wiki? Large wiki systems like wikipedia and emacswiki function just fine with the "anyone can edit" policy. There is already a huge 2+2 community, why not take advantage of that?

[/ QUOTE ]

It sounds like you've got even grader ideas than I have (which isn't bad). The problem with teaching pages is that we then have to make sure that the teaching is right, make sure nobody screws with them, etc. By having just a catalog, we avoid those problems.

But the wiki idea may still work for a catalog. I've not played with the wikipedia, so I don't really know much about it. How hard is it to set something up? And what are the search capabilities and organizational structure?