I feel really weird just posting this. In fact, even if you think I'm just an idiot, feel free to post that. It's actually good feedback.

I read the FAQ at http://forumserver.twoplustwo.com/56.../msnl-faq-182/, and it seems this is okay for a thread. It's not in any of the don'ts. It's definitely strategy related, and definitely about mid-stakes.

And yes, there is something in it for you, personally, MSNL regulars, in reading and responding to this post.

What the hell are you talking about, guyupstairs?

Ok, here goes. I've written some training software. It's called Range vs. Range, and it's live now, and free. I've already had people say it has the potential to make the games tougher (to be specific, one $10-$20 LHE pro, and one 200NL regular have said this). It may or may not be true, but it's some evidence.

(In case it's not obvious from the below description, the software is about improving your play against regulars, so that you can better exploit them, but also to be used defensively so that you can move up to where the fish have more money.)

Well, I've recently had an idea for how to significantly improve this training concept. (Skip to the end of this section if you're not interested in the details.) In short, it's playing poker, or playing a training situation as in the current Range vs. Range system, except instead of playing your hand in the context of each player's range and the hand you are dealt, you will now be playing your entire range. Instead of the server asking you what action you want to take with your current hand, the server asks you what action you want to take with each hand in your current range. You will be asked to split your current range into a folding range, a calling range, and a raising range (and specify a raise size). More details on this to follow if you're interested.

This training concept can also be applied to a full hand (heads-up, 6-max or whatever), or within the context of a training situation such as the current ones defined in the existing Range vs. Range software (BTN calls MP's open; BB 3bet and CO call; BB squeezes and gets called by both players; etc., etc.).

Well obviously this is intended to teach range-based, near-optimal play, from which to adjust based on observed weakness in your opponents. If you get good at this, playing hands at the poker table will be easy. After every hand both players will be told about all the range selections of their opponent(s), so that they can quickly adjust towards more optimal play. If there's a showdown, the pot will be divided based on range-vs.-range equity, using the final ranges for each player. Results will be recorded as they currently are for Range vs. Range, and can be compared to the average of all players (called EV in that software).

Good for you. You wrote some software. This isn't the software forum, it's MSNL! We discuss poker.

Yeah, ok, here's the point. I think there's some chance that my training software will make it significantly easier to learn NL HE strategy, and make the games significantly harder. But I do have a potential solution.

Firstly, here's what I consider the possibilities to be, with respect of the usefulness of this software:

1. It won't actually work as described. It's creating unrealistic situations, and so can't teach you anything. You're not actually practicing the skills you need at the poker table.
2. It's effectively useless, or a curiosity at best. Sure you can use it to train, but the skills really aren't transferrable to real poker.
3. It's useful in a small way that makes it worthwhile using, but it's not sufficient to compete with other training methods. You're better off just reviewing your hands, or watching a video.
4. It's sufficiently useful to compete seriously with other methods, but not enough to make the games significantly tougher. It's just one more option, useful to people who don't like the other tried and tested training methods available to them.
5. It's good enough to make the games significantly tougher. The learning curve to start meaningful training with the system is sufficiently small (especially augmented by the existing range-vs-range training system which lets you play a single hand at a time) that many people will use it, and get significantly better at the game.
6. It's a killer app, like Sit'n'Go Wizard. It becomes the de facto standard in perfecting your ring game play. The games get much tougher, higher stakes become accessible to players who have been stuck at lower stakes. Bankroll becomes just about the only differentiating quality between lower-stakes and higher-stakes players.

Sure, they're all at least conceivable. Frankly I think you're just a crackpot weirdo.

I'm almost done. Having read this far, you may as well hear me out.

I find possibilities 1 and 6 unlikely, though I can't rule either of them out entirely. In the case of possibility 5 or 6 though, we collectively have a problem. I don't want to make the games tougher. What I would prefer, in either of these cases, is to make my software exclusive, by charging enough for it that few people get to use it. Perhaps having limited subscriptions available, and auctioning them off. The problem with this is that there would be nothing stopping someone else from writing their own version of this software, and giving it away for free. That would be something that nobody wants.

So I have applied for a patent for this idea. (Yes, this kind of thing is patentable. I looked into it.) If the conclusion is that possibility 5 or 6 is the case, then I can continue to patent the software and make it exclusive, stopping most people from getting the benefit of this software. If the conclusion is that possibility 1, 2, 3 or 4 is the case, then I will let my patent application lapse, the concept will go into the public domain, and I will continue to develop my software and make it freely available to all (in cases 3 and 4, at least). I won't be able to charge any significant amount for it, because then someone else could just make a cheaper version of the same thing.

So here's my question to all of you. Which of these possibilities do you think this new training software concept is? If the answer is 5 or 6, it will be your benefit to tell me, so that I continue to patent this concept, so that it doesn't become effectively free to all.

You arrogant dick. You really think you're that good?!

No, I just think it's possible that the idea I came up with has that much potential, and I want to find out one way or the other before I get back to doing what I'm really interested in: finishing the software.

it's been a long night of partying but you lost me at where the software asks me what I want to do with my range here instead of vs my actual hand... except when we have a good hand (although it's a small part of our range there) Then wat?

god, backward. How about dont develop the software to begin with? This is how botting came about.

But do you think that's necessary? That's the question this thread is about.
What does this mean? I don't understand. Are you saying the bot writers shouldn't have developed the botting software? Well, perhaps, but that's like saying criminals shouldn't break laws. They shouldn't, but they just don't care, and there's money to be made.

This type of software has been around for a while. Its called Cardrunners EV/Flopzilla.

Is it? How do Cardrunners EV and Flopzilla tell you whether or not your strategy is exploitable? That's a very important question if you want to do better against regs, and move up. But I have to admit that I don't know much about those tools.

But with Range vs. Range (the current version, and the new planned version), you are actually playing poker against another person the whole time. With this new concept, you will be forced to decide how to play every hand in your range at every decision point, as you play your hand, without knowing what your hole cards are. That's a skill that is instantly transferable to playing poker at the tables. And after every hand, you will get to see what your opponent chose for their ranges at every point (potentially; it'll be an option), so you can see how they are attempting to exploit you (if at all).

Edit: Sorry, I shouldn't be defensive, even if you are being dismissive. I should accept your reply as what it is: a vote for option 4 (possibly 3). I'm happy to go ahead and make this software, and give it away for free, if that ends up being the case. Thank you. I do appreciate your feedback.

dont you get a huge, probably unmanageable tree if you want to have each hand in each range reach the end of the hand

You don't have every hand reach the end of the hand. You deal a hand to each player, but don't tell them. Each time it is a player's turn to act, they carve up their residual range (i.e. their range from their previous action) into three new ranges (fold, call, raise, and also a raise size), and tell the server. Then the server decides which action to take based on the hand they happened to have (this is why no tree develops). The player never gets to see their cards though, so they always have to decide what to do with their entire range. If they get to showdown, instead of revealing their true cards, the server can just apportion the pot based on range vs. range pot equity using their residual ranges.

Edit: I'm off to bed. I'll be back in the morning.

When I read your initial thread I forwarded it to a friend under the heading 'the end of online poker' not in it's current form but I can imagine if a future version of this becomes widely incorporated into huds or bots it'll screw things up. In reality though people are too lazy or stupid in general (including myself) to either see its value or use it to its potential. Are you asking for hush money in a roundabout way? Charge a fortune for it and best of luck.

But if you don't let all hands in all the ranges "finish" how are you going to get their fair share of the pot / in EV? E.g., doesn't what you outlined favor slow playing vs a bet as most likely I will have a hand in the calling (or folding) range?


I don't fully understand how you plan to implement this, but in the ways I can imagine this doesn't work out / doesn't return conclusive results.

Yeah, there seem to be a lot of hurdles to making this software work. I think the feedback to the players and evaluating the meaning of the equity calculations (ie what strategic changes players should make) will make or break this software. I can really see any of the outcomes above happening, although I think 2 and 4 are slightly less likely than the others.


This has always been my unfortunate experience with CREV. Too much work for not enough information. I'd really like any kind of alternative and I find the idea of this software really intriguing.

This is not CREV.

It seems an example is in order.

Two people are practicing the "MP open; BTN calls" situation. (Alternatively, we could reach this starting position by playing a 6-max ring game starting from the very beginning of the hand, in which case the players get to choose their own opening and flatting ranges.) MP's opening range is {66+,A2s+,KTs+,QTs+,J9s+,T9s,ATo+,KJo+}. BTN's starting opening range is {TT-22,AJs-A2s,KJs+,QJs,JTs,T9s,98s,AQo-AJo,KQo}.

The server deals a hand to each player, but doesn't tell them what hand. I'm not even going to bother giving an example.

The flop is Kh 8d 3c. It is MP's turn to act. MP decides to cbet {KK+,88,A2s+,QTs+,ATo+} and check {QQ-99,77-66,KTs+,J9s+,T9s,KJo+}. MP's true hand happens to be in his cbetting range, so the server performs his cbet. MP's range is now {KK+,88,A2s+,QTs+,ATo+}.

BTN decides to call the cbet with {TT-22,AJs-A8s,A3s,KJs+,98s,AQo-AJo,KQo}, raise with no hands at all, and fold with everything else. BTN's true hand happens to be in this calling range, so the server performs his call.

The turn is 2c. MP bets {KK+,88,AKs,A8s,A3s-A2s,ATo+} and checks {AQs-A9s,A7s-A4s,QTs+}. The server recognises that MP's true hand is in the betting range, so it performs the bet. Similarly BTN's call range is {TT-88,33,AJs,A8s,A3s,KJs+,98s,KQo}, which includes his true hand, so BTN calls.

The river is 2s. MP bets {KK+,88,AKs,A3s-A2s,AKo,ATo}, BTN calls with {KQo,KQs,88,33}.

There is a showdown. We still don't know what either player's actual hand was, but we don't care. But we do know that:

• If MP had known what his hole cards were, he would have played exactly the same way.
• If BTN had known what his hole cards were, he would have played exactly the same way.
• MP would have wanted to play every other hand in his final range exactly the same way, and so would BTN.
• Neither player would have wanted to play any hand that is not in their final range like this.

So all that matters is that the players reached a showdown with these ranges. The server performs a pot equity calculation, BTN's range has 59% equity, so BTN gets 118.5 big blinds, MP gets 83 big blinds.

Then there are some options. The server may tell each player the other's range. The server may tell each player what range decisions the other made for each action. The server could keep both players in the dark, and even just show the hole cards and give the pot to whoever had the best hand, although I don't think that would be as useful for polishing your strategy. But in any case, it's up to you to do the work of changing your ranges to exploit your opponent based on what you've seen of them. The software is not going to tell you how to play, no way, no how. That's still your job.

Too much work? I don't think so. At least as long as I come up with a really good UI for splitting your current range in to three sub-ranges.

I didn't say or think it was. Internet brought CREV up and I agreed that the other software options for exploring optimal play are not very useful/easy to use. Whether or not your software is like CREV it definitely competes within the same space. I like the idea of your software and it excites me - I would definitely try it.

As for your example, I'm not sure why the software has to deal an actual hand at all. Why not just explore range vs range play without assigning a specific hand to each player? As for what sort of information to give the players at the end of the hand I think it's best to give as much as possible, but I'm not sure what format it should be in. It's useful to know villain's rangeplay and our equity against it, and having some kind of tool to explore strategy options and their value would be very useful.

are you a bot

Pm me for a HU game with your magic software