Website: https://deepsolver.com/
Pricing: Start with Free Trial. Upgrade to paid version ranging from $10 to $209 /month (up to 40% Early Bird promo included)
Contacts: Discord - https://discord.gg/Nt8h2M5Nqj , Email - support@deepsolver.com
Get 50% extra coins for first month: Use this link when creating an account



WHY DEEPSOLVER?

Deepsolver is the fastest poker solver and customizable GTO trainer in the world, delivered in the Software as a Service (SaaS) model directly to your browser, with subscriptions ranging from $10 - $209/ monthly. Deepsolver solves two-player post-flop No-Limit Hold'em, with multiway pots and many other formats (such as Pot-Limit Omaha and NLH Short-Deck) to come in the near future.

The new app cuts solving calculations times to just a few seconds, making it the fastest poker solver in the world. Combined with a real-time, customizable GTO Trainer, Deepsolver offers poker professionals unprecedented speed and convenience of use. Thanks to the use of cloud computing and artificial neural networks, Deepsolver users can learn GTO strategy any place, any time they like. No need for expensive hardware, updates or costly setup and maintenance.

Deepsolver is available to purchase in GTO Trainer + Poker Solver Bundle in the following subscription plans:

• Recreational ($10/monthly) - GTO Trainer only. For passionate enthusiasts starting their adventure with GTO. Poker is not your entire life, but you would like to advance your understanding of GTO strategy in a fun way.
• Regular ($76/monthly) - GTO Trainer and Poker Solver in one bundle. For semi-professional and professional poker players, wanting to raise their edge over the competition. You earn money at the poker tables and want to earn more.
• GTO Pro ($209/monthly) - for professional players and GTO veterans, using solvers and GTO Trainer as part of their daily warm-up and cooldown routines. You run lots of calculations and analyze your game daily.

FEATURES:

Fastest in the World



Our Neural Net makes Deepsolver incredibly fast. So fast in fact, that we were able to cut calculation time from minutes to seconds or even hours to minutes.

What is unique about Deepsolver is that we mix the traditional CFR algorithm (proprietary one) with the power of artificial neural networks and freedom of digital cloud to deliver an unprecedented speed of calculations and seamless user experience for poker professionals.


360° GTO Trainer



With our GTO Training Mode, you are able to set up custom boards, ranges, pots, stacks and any scenario you want. Train against the best player in the world (our AI) and crush your stakes with ease.

No need to buy $10.000 USD servers, no need for hundreds of GB of RAM. Deepsolver is sold in a subscription model with different price points, adjusted to the need of players at different stakes and playing various game types. Enjoy access to your poker solver any time, any place in the world. All you need is a laptop, internet connection and browser.

Hand Locking



With Deepsolver you can lock your oppositions range - as you know their personal tendencies - and prepare optimal counterstrategies for any opposition at any level. Poker is a game of thin margins and very sharp edges - raise yours with our next generation poker solver.

Advanced Analytics



Learn fast, learn easy and most of all - learn how to win - with our advanced Analytics and Visualisation modules. You'll get your results in easy to understand format, with the most complete assortment of tools available on the market. We are eager to implement new features our users request.

Multitasking

We designed Deepsolver from scratch with multitasking in mind. With our user-friendly interface, you can easily, seamlessly switch between tabs and processes. Each solution, scenario or even draft is saved and stored for future use and convenience in your own private easy to access database.

Fair play

Many of us are former professional players so we at Deepsolver fully understand the growing concerns over the assisted use of solvers during online play. We also absolutely understand the worries of the poker community whenever new software comes out that offers fast solutions. We treat these matters very seriously so we have taken measures to counteract such potential activities using Deepsolver. It is also worth noting that each hand that is solved using Deepsolver is stored in our database so that in the case that a player is suspected e.g. by a poker site we are more than happy to cooperate and can cross-reference his or her hand histories with our own to see if there are matches or striking similarities. We love poker, want to keep the games clean and will do whatever we can to help with that.

Imagined by Pros | Designed for Pros

Deepsolver is founded and imagined by professional poker players, who brought out together a team of developers, AI engineers, data scientists and designers to redefine the experience of poker solvers. To change the game.

That claim and promise extend to the quality of service and support, as well as long-term technical support and development commitment. We are here for the long-term game.

Did anyone use it? What do you think about this solver.

I'm using deepsolver and recommended it to my students that has not used solvers before, it's very fast and convenient. Huge fan of this project!

I'm personally involved full-time with this project (have played poker for 16 years and in the last 4 primarily 25/50 and 50/100). I think we have something pretty unique here and I don't think it's a marketing gimmick or a stretch when we call ourselves 'next-generation'. Thanks to our hybrid of standard CFR algorithms and neural networks we can cut down sims that take single or double-digit minutes in any other solver to seconds in Deepsolver. We're web based so you don't have to spend 10k$ on a decent server or worry about how and where to maintain it. We're sort of the best of both worlds of solvers and library based apps, because we offer full customization (in terms of bet sizes/stacks etc.), nodelocking, new/unique features down the line on the one hand and on the other - ease, convenience, accessibility. We're also explicitly against RTA and genuinely will do our best to flag such activities and cooperate with sites to weed out the bad apples - not doing it just for PR points, would probably be more profitable to say nothing about it and get some 'pretty active' subscribers instead.

I can also honestly say we have a lovely team of really cool, talented people who are working on the app everyday and making incremental improvements literally each week. We also have several former high-stakes players invested in this venture and giving insights on a regular basis.

Lastly, as a former high-stakes HU PLO player, I feel very excited about Deepsolver supporting Omaha hopefully somewhere around midway this year, because with Omaha our difference in solving speed will be even more pronounced and also by then we will have a pretty long list of features (some entirely unique) to go along with it.

When you say multiway pots do you mean it can also solve 3way postflop spots?

Multiway postflop spots are on our roadmap (and it will be really cool to cut down calculation times for those using our technology, because currently they take ages), but Omaha will come first. Realistically multiway should be supported by the end of this year or very early 2023.

ok ty

IIRC others solvers devs claimed that 3way+ equilibrium is very shaky due to different incentives players might have. I also recall a Cardrunners video where the coach demonstrate that 2 people playing GTO against a third player who play suboptimally, will lose. How do you overcome this and what do you think about that?

Do you remember the video's title or who made it? If it was a Cardrunners video I assume they were using Pokersnowie at the time?

As for different incentives - do you mean cooperating with other players?

It was made with CREV
Regarding the coach, take it with a grain of salt, but it should be the creator or GTO Range Builder

Ok, found it (was also in blog form), was a very interesting read.

In his example 3 people were playing a push/fold game. These are their winrates with all of them playing GTO:

BTN EV: 19bb/100
SB(hero) EV: -11bb/100
BB EV: -8bb/100

If we set BTN to play tighter than GTO (subotpimally as you put it) these are the new winrates:
BTN EV: 15bb/100
SB (hero) EV: - 17 bb/100
BB EV: 2 bb/100

BTN naturally lost EV, BB gained EV, but what is interesting is that SB lost EV. And mind you this is not collusion, it's just 1 player deciding to play a bit differently.

The definition of Nash equilibrium (GTO) is a set of strategies where neither player can increase their EV unilaterally (a player cannot increase their EV relative to what it is in GTO). A lot of people have come to assume that means another player can't _decrease_ our EV compared to what it is in GTO, but that's only true in a 2 player environment (one person's gain is another person's loss). In multiway spots another player can definitely decrease our EV, but the definition of GTO holds - they can't increase their own EV (unless they are colluding and profit sharing with the remaining player).

It's different from the way you phrased it ("2 people playing GTO against a third player who play suboptimally will lose"). BTN lost 4bb/100, SB lost 6bb/100, but BB gained 10bb/100, so just 1 person was hurt here by BTN's supoptimal play. Assuming everyone gets an equal amount of hands on each position (BTN,SB and BB) you will actually not lose EV in the longrun.

I'd say for multiway solvers still serve their purpose, but perhaps there's more utility in nodelocking there than compared to heads up situations.

Glad that you have found it useful (care to share the blog post you have found, because I haven't)

BIG CHANGE TO THE COIN SYSTEM

We are simplifying the current coin system. We understand that the in-app pricing that we had in place thus far could have been confusing to many. While playing a hand in GTO Trainer was worth 2 coins, solving flops in Custom Solution was worth 10 (with turn and river being free of charge). Since Bulk Calculations are coming very soon we would have been forced to create a separate pricing model for that as well. So we decided to make all of this easier for you guys.

Starting today solving 1 street in any feature = 1 calculation.

Apart from making the system more comprehensible, in essence Deepsolver just got a LOT cheaper - you can now run up to 10x as many Custom Solutions on Regular and GTO Pro Tiers compared to the previous system.

Happy solving!

AGGREGATED REPORTS ARE HERE

Today we're very happy to announce that we're finally releasing aggregated reports (bulk flop calculations and turn/river runouts).



Within the Bulk Calculation feature you will find board texture packs - mixed (if you want to take a global look at various textures at once) and specific (if you'd prefer to focus on a single type of board e.g. mono).



As always you're free to choose any parameter you want (range,stack,sizing) and have dozens of boards all calculated live faster than you'll be able to make coffee.

Enjoy!

I'm using it on my free trial period and I have questions regarding solutions. Some solutions seem slightly different from GTO wizard solutions. For example, on AK8ss board I build tree using one sizing 150% and solver wants to overbet AA 100%. Doesn't really make sense to me. Shouldn't we check it back most of the time? Then I've looked up same spot in GTOwizard and there it says you mostly check it back and when you bet, you bet small. So which one is it? Why deepsolver solution is different?

Trainer was painfully slow. Took like four seconds after every decision to move on. Did not like that. Been looking for one and tried pretty much everyone mentioned 2+2 and this was the slowest one with pretty big margin. Allso don't like that coins system. Just one ammount and x ammount of solves you can do. No need to invent the wheel again. I mean 500 calculations you might think it's 500 different games but no. 500 different streets. On more positive side the interface is very nice and easy to use really liked that.

A common theme we see is our solver choosing a bit more polar betting strats (bigger sizings / less slowplays). Betting strategies might differ from other solvers , possibly because we use a unique technology (CFR+neural networks instead of just CFR) and there are various ways to achieve GTO for a betting distribution. We have done a variety of accuracy benchmarks ( https://deepsolver.com/blog/speed-pr...ks-and-testing ) using different boards/SPRs/ranges (720 spots) and had a mean ND of 0.59% (checked by locking strategies into pure CFR solvers) so the strategies should be solid even though they might look weird.

We will be making a big update in the coming days which includes, among other things, a multitabling GTO Trainer (up to 4 tables) and loops - allowing to replay the board/street/node/decision if you see one of interest to you. The play within the loop is instant as you are operating on an already calculated sim.

The 4 second delay you were experiencing is due to the fact that Deepsolver has the only completely live trainer in the world - each street is calculated before your eyes. However right now you can get an instant experience by calculating a sim through 'Custom solution' and using the 'Play vs. solution' feature. You mentioned trying other trainers on the market - you might find Deepsolver preferable (in the instant 'Play vs. solution'or upcoming 'Loop' mode) because you are not limited to the set of libraries provided by the apps. Here you can create any scenario (stack/range/sizings), calculate it in 10 seconds and start practicing it without any delays. You can also do that for a sim you nodelocked using our app which is very appealing to a lot of players.

As for the 1 street = 1 calculation system I understand what you're saying, but its in place, because our solver calculates each street independently and given how that works it would be very difficult to come up with something that is more clear and comprehensible.

We're happy to hear you enjoy the interface.

We have just released a HUGE update for Deepsolver and we’re more than excited to present to you how our app has improved. With this update you get:

🔔 New features:

Multitable GTO Trainer:

- Choose between two different modes “Challenge” and “Educational”
- “Challenge” - up to 4 tables, fast-paced environment that simulates a real online poker session, perfect to test your skills or as a part of your warm-up routine
- “Educational” - 1 table, a great choice for an in-depth learning experience with the assistance of many helpful tips
- “Loop” - a spot came up on the table that you find interesting? “Loop” allows you to replay that specific situation - board, street, node or even just the decision


Simplify:

- A solution suggests a strategy with low-frequency bets? “Simplify” is a feature that recalculates the sim without rare bets just by pressing a single button!
- No need to waste time manually deleting uncommon sizings yourself
- You choose what is the upper limit of the frequency to be removed


Multi-Street Tree Builder:

- Customizable bet sizings for each street (flop/turn/river) in GTO Trainer
- Gives you full control over the decision tree
- “Apply to all” option if you prefer fixed sizings for every street


📈 Improvements

- Engine accuracy improvements
- Cloud improvements
- Mistakes and blunders are displayed through EV loss
- Bets and raises can now be set to display as percentages in “Preferences”
- Solutions should now load faster
- Crypto payment is available
- Solutions are now automatically named by their position, board, SPR
- We removed pre-flop ranges that were extremely low in frequency

🔧 Bugfixes

- Fixed an issuse with coinpacks
- Fixed a bug where marketing rights terms were automatically accepted
- Fixed bugs related to incorrectly displayed fonts
- Fixed a bug where GTO scores weren’t updated before the end of the hand

➡️ Wanna know more? Go to our website and see the latest blog note: https://deepsolver.com/blog/deepsolv...y-training-app

➡️ Any feedback, questions or other issues? Contact us: support@deepsolver.com.

I have a question.The accurancy you run the sims compared to other sites is lower for example 0.5 to 0.2 that most of the other sites with presolved sims have.Do you think this doesnt impact the solution or do you plan to improve the accurancy of the solutions?

Hi,

Right now we average 0.59% ND based on the testing we did (720 spots with various boards/stacks/ranges). An AI with <2% ND would easily crush the best players in the world so with 0.59% you're getting outputs that are clearly unbeatable in any game. That number is a bit worse than what sites with presolved sims offer, but with Deepsolver you can do things that you will never be able to do with them - hand lock, edit the pre-flop ranges, pick whatever sizings and stack sizes you like, change the rake structure. That said, our team has noticed some things that can be improved in the engine so we are planning on significantly improving our accuracy some time later this year.

Recently we have updated our 💸 monthly subscription plans 💸 so now you get:


♣️ GTO Pro - UNLIMITED calculations (exclusive pack included)

♣️ Regular - 5x more calculations than before, the same price

♣️ Recreational - 1000 streets to calculate for only 19$ per month


Choose the best plan for you and elevate your poker strategy with Deepsolver!


➡️ Claim up to 40% promo: https://deepsolver.com/pricing

➡️ Any questions or issues? Contact us: support@deepsolver.com.

With all the scandals going on we understand the growing concerns about cheating. Deepsolver has been explicitly against RTA from the start and now we have created a tool for you to investigate how we are being used: https://deepsolver.com/fair-play-check

This tool allows you to check if and when a board was solved using Deepsolver so you can compare the findings with hands that were played at the same time online.

We hope steps like these inspire other poker apps to follow suit and create similar ways to keep themselves accountable. There's been a lot of dirt uncovered lately in the poker world. Perhaps together we can take a stand and make something good out of this?

APRIL'S UPDATE TIME

🔔 FEATURES:

- Customization - you can now change the look of your cards, choose colors for actions and edit the font size



- Fair Play Check - a new anti-RTA tool that allows the community to investigate if and when a board was calculated using Deepsolver



- Onboarding - new users are greeted with an onboarding and a suitable tutorial



📈 IMPROVEMENTS:

- !New pricing! The regular tier has 5x the calculations, Pro is unlimited!
- The GTOTrainer view has a fresh new look (table and animations)

🔧 BUGFIXES:

- Fixed a bug in Challenge mode
- Fixed a bug where Gutshot and Combodraw were not displayed in the Hand strength module
- Fixed a bug related to positions when changing streets
- Fixed a bug where the Pocket pair below the top was showing 2 hands max
- Fixed a bug where after node locking nodes that had 0% frequency were still available
- Fixed a bug where chips/sizes were hidden behind a players label
- Fixed a minor issue with Aggregated Flop Report

Is GTO traier unlimited even for basic and pro t iers?

Hi. GTO Trainer is unlimited only on GTO PRO tier. On Recreational you'll get 1.000 calculations, while on Regular 5.000 calculations (1 calculation = 1 street).