Methodology
By Theo Brand, independent poker software enthusiast · hello@bestpokerbots.com
Eight poker bots, 45+ criteria, one composite score out of 10 each. Everything scored here comes from what the vendor publishes about itself, read in October 2026 and saved. Below is exactly how the score is built, what it deliberately ignores, and what this ranking is not.
What the score measures
It measures how well a product serves a buyer who wants to know three things before paying: does it work in my room, what will it cost me, and where does it run.
It does not measure profit. No bot in this category has a win rate that anyone outside the vendor has verified, so a score that claimed to predict results would be inventing them.
Where the data comes from
- Vendor pages. Room lists, format tables, price lists, platform statements, server locations, support arrangements. Read in October 2026, with copies kept so an entry can be checked against what the vendor said on the day it was written.
- Vendor pricing material. Published tariffs, calculators and package lists where they exist. Four of eight publish no price at all.
- Operator experience. The author runs this kind of software on his own accounts. That shapes which questions the ranking asks — per-room formats rather than platform counts, cost per hand rather than headline packages — but it is not presented as measurement, and no private result is used as evidence.
Nothing is taken from affiliate feeds, press releases written for publication, or forum claims that cannot be traced to a vendor page.
The six dimensions and their weights
| Dimension | Weight | What earns points |
|---|---|---|
| Room coverage | 30% | Rooms named individually, formats listed per room, table-selection support, tournament coverage. A platform count with no breakdown earns almost nothing. |
| Published pricing | 25% | A per-hand tariff, a calculator, or packages you can read before contacting anyone. A licence quoted only in private chat scores low. |
| Vendor disclosure | 20% | Operating history, engine description, results with a stated sample size, server locations, support arrangements, language coverage. |
| Platform honesty | 15% | What actually runs where: native Android, native Windows, or an emulator layer the vendor would rather not mention. |
| Free testing | 10% | A demo, a trial, or a cheap ticket that buys real hands rather than a slideshow. |
A missing fact is recorded as not published and scores zero for that criterion. It is never filled in with an estimate.
The scale
- 9.5–10 — publishes everything a buyer needs before paying: rooms, formats per room, a price you can calculate, platform, infrastructure, and a way to test.
- 8.5–9.4 — strong on coverage or documentation, with one structural gap, usually the price.
- 7.5–8.4 — a real product with a real technical position, but several things you cannot check until you are in a conversation with sales.
- Below 7.5 — little published, or nothing an individual player can buy.
Club-side products, sold to the house rather than to the player, are scored on the same criteria but cannot outrank a player-facing product. A reader cannot buy them.
Ties
Where two products land within 0.2 of each other, the one that publishes more about itself goes higher. Where they still tie, the one that can be tested for free goes higher.
What is deliberately not scored
- Anti-detection claims. Several vendors publish checklists about staying invisible to room security. Those claims cannot be verified from outside, and ranking them would amount to publishing a guide to evasion. They are not scored, and the details are not reproduced anywhere on this site.
- Vendor win rates. Reported in the entries, because readers want to see what is being claimed, and never converted into points.
- Design, marketing and audience size. None of it tells you whether the bot runs in your club app.
What this ranking is not
- It is not a bench test. No controlled test was run for this edition, no hand sample is published, and no vendor was asked for a test licence. Anyone publishing a ranking of this category with verified win rates owes you their dataset.
- It is not advice to buy. Third-party assistance breaks the terms of service of most rooms. The account, the balance and the consequences are yours. See the terms of use.
- It is not paid for. No vendor pays to appear, to rank higher, or to be removed, and no entry is shown to a vendor before publication. Outbound links are plain
nofollowreferences so a reader can check the source.
Known limitations
- Vendor pages change without notice. A price or a room list can move the week after an edition is published. The date at the top of the ranking says when the facts were read.
- Unverifiable claims stay unverifiable. Hand counts, win rates and accuracy figures are statements, not measurements, no matter how precise they look.
- One person writes this. There is no second reviewer. Mistakes get fixed when they are reported, which is why corrections with a source are handled first.
Updates and corrections
The ranking is dated. When a vendor changes something material — a price, a room list, a product line — the entry is rewritten and the date moves. Changes between editions are listed in the What changed in this edition section of the ranking.
Found an error? Write to hello@bestpokerbots.com with a link to the page that proves it, or use the contact form. See also the editorial policy and the ranking policy.
← Back to the ranking