You can see the balls. What is hard to see is which line is actually the best one, how thick to hit it, and where the cue ball ends up afterwards. AimCue simulates the table you just photographed and shows you the shots a real player would consider, with the odds attached.
Take a picture of the table, confirm the four corners, and drag any ball that landed in the wrong place. The confirmation step is the point: you are the one who says the layout is right before anything is calculated.
Touch a ball on the table and that becomes the target. AimCue answers the question you actually asked, rather than picking a shot for you and leaving you to work out why.
Every suggestion names where to aim on the object ball, where the tip meets the cue ball, how hard to hit, and where the second contact lands. The tip diagram is drawn from the same numbers the simulation used.
Libre, four-ball, three-cushion, 8-ball, 9-ball and snooker. Carom and pocket tables use different geometry and different solvers, not one engine pretending to cover both.
Eighteen drills with a position you load onto the table, a short lesson, self-scored reps and a day streak. The daily plan rotates so practice is decided before you pick up the cue.
A summary of the international rulebook for all six disciplines, drawn from the UMB, WPA and WPBSA texts, so a table argument ends in the app rather than in an argument.
AimCue runs the Mathavan 2010 cushion model, ported equation by equation from an open-source implementation of high-speed camera measurements of real cushions. It models the second contact point, where the ball is pressed into the slate during the impact, which is what shortens a shallow track on a real table.
Two other cushion models, Han 2005 and Stronge, are kept in the engine purely so the port can be checked against pooltool, the open-source physics reference. Each passes 600 of 600 golden fixtures, agreeing to floating-point precision rather than merely within tolerance.
A shot is not scored by a formula. Each candidate is played out 48 times with the hand noise of a real player at your stated level, and the number you see is the share of those attempts that still scored. This is the method used by the billiards programs that have won the Computer Olympiad.
Tip placement is capped at the miscue limit, enforced in the search, in the refinement and in the noise model alike. A cue tip more than about half a ball radius off centre slips, so AimCue will not propose the enormous swerving lines that come from spin nobody can actually apply.
No. AimCue makes no network requests at all. There is no account, no server, no analytics and no third-party SDK, and the app contains no networking code. The physics simulation and the computer vision both run on your device, so a photograph of your table is processed and stays there.
Six: libre, four-ball and three-cushion on a carom table, and 8-ball, 9-ball and snooker on a pocket table. The pocket disciplines use a separate solver, because on a pot the geometry fixes the aiming line and you choose the force and the tip, whereas on a carom the thickness of the hit is yours to choose.
By replaying the shot. Each candidate is simulated 48 times with random error added to the aim, the force and the tip position, sized to the skill level you picked, and the percentage is the share of those replays that still scored. A club-level profile uses a tighter spread than a beginner profile, so the same shot reads differently depending on who is holding the cue.
It is designed not to, and this is covered by tests: every three-cushion suggestion is checked to be a legal carom, and four-ball advice is checked never to send the cue ball into the opponent's cue ball. If you tap a ball you are not on, AimCue still answers with a shot you are allowed to play. It advises on the shot in front of you and does not referee your match.
Honestly, not yet measured to a standard worth quoting. It locates the table in most photographs, and it is weaker when the picture is taken from along a long rail. Because of that the app is built around you confirming the four corners and correcting any ball, so the layout that gets simulated is the one you agreed to rather than the one a detector guessed.
The engine carries carom tables at 2.84 m and 2.54 m, a 9 ft pool table and a 12 ft snooker table, and the cushion parameters for carom are calibrated against the five-and-a-half diamond system rather than left at a library default.
Not yet. It is finished enough to play with but it is not on the App Store, and this page will carry the download link the day it is. Leave your email and we will tell you when that happens.
Yes. The free tier gives you a limited number of solves each day, one suggested line per solve, the first few drills and a short history. AimCue Pro removes the daily limit, shows every option it found, adds position play for the next shot, and unlocks the full drill library and history. Pricing will be confirmed when the app reaches the App Store.
AimCue is a training and analysis tool for cue sports. Suggested shots are the output of a physics simulation of the layout you confirmed, and a success percentage describes simulated attempts at the skill level you selected, not a promise about your next shot. AimCue is not affiliated with the UMB, the WPA or the WPBSA; its rule summaries are for reference and the governing body's published rules always take precedence.