Hi,
(This is my first post in lifein19x19. dokidoki^^)

I have developed and released an iPad app, which uses AlphaGo Zero type engine with ELF OpenGo weight (v1).
Here is an introduction page of the app.
https://new3rs.github.io/a_master_of_go/index.html

Brand-new A12X Bionic processor achieves max. 170 playouts per second. I think that this value is better than usual note PCs' and comparable to usual desktop PCs' though not better than gaming PCs'.

I hope that people here will be interested in my app.

Thanks!

I just bought it; the price is €10,99. The app's name is "A Master of Go", the Japanese name is "囲碁の師匠" ("Igo no Shishou").

Having ELFv1 in a Lizzie-style interface on an iPad is very nice. On my 2017 iPad (9.7", not Pro) with an A10 processor I get between 9 and 10 playouts per second, which is not too bad, given the convenience.

There is a mode where you can play the AI or watch it play itself. Then there is a study mode where you can explore positions starting with an empty board. You can also paste SGF for play or study modes.

When starting a game in play mode, you can set the number of seconds per move, whether to save battery by not pondering during the user's thinking time and how many threads to use. The komi is fixed at 7.5.

In play mode, after resigning a game, you're in study mode so you can explore variations.

In the Dropbox app, you can export an SGF game to the app.

Some issues:

- During "play" mode, tapping the "Resign" button immediately resigns. So even an accidental tap ends the game. I would prefer to have a confirmation dialog ("Do you want to resign?").

- During AI self-play, the "Pass" button in the bottom right is visible but doesn't do anything. The "Resign" button however, as expected, ends the game and goes into study mode.

- The win rate graph above the board is not very clear, with yellow and green lines on a grey background. In a position with 55% I still only see an almost flat line. Something closer to Lizzie would be good, including the black/white bar graphically showing the current win rate and how much the last move gained or lost.

- When going from play mode into study mode, the win rate is completely flat; that is, I think it only updates the win rate when a move is made.

- There is a button with numbers underneath the bottom left corner of the board; tapping it toggles the button, but I don't know what it does.

- It would be very useful to have a variation tree display that you can navigate.

- It would also be nice to be able to save games to Dropbox or Files.

- The controls for navigating back and forth show a double left triangle for going back and a single right triangle for going forward. Why?

- In the "Start game" dialogue, it says "Play in Black" and "Play in White"; more natural would be "Play as Black" and "Play as White". Also "idle during user is thinking" might be better worded "idle while the user is thinking".

Thank you for developing this app; I'm looking forward to both using it and seeing how you continue to develop it.

I bought it to review my pandanet games but the app it's unable to load pandanet sgfs , maybe the nod count can be against a solid color to better appreciate the number? , overall an interesting app , hope to see some improvements in the future.

Grünauer-san,
Thank you so much to purchase my app!

I promise to improve the issues which you kindly pointed out.

About the button with numbers, it means a submode to show variations which the AI is reading.
While the button is highlight, if you touch a candidate move, then the primary variation of that candidate move will appear such as mouse-over on Lizzie.

You encourage me to develop the app very much.
Thank you!

Kaliwan-san,

Thank you very much to purchase my app!
I will research pandanet SGFs.

I will never notice this issue if you had not told me.
Thank you!

to be able to disable sound will be good and icloud integration for loading sgfs will be a nice feature

Import from kgs does not work.

Can someone explain the meaning of the different colors and the red circle?

I see; that's very useful.

However, there is a bug. See the attached screenshot, which has multiple moves 4 and 5. I've let it run for a while; more moves 4 appeared. Is it possible that when a new variation appears the old one is not cleared?

I love this--there's a real gap for strong AIs in the iOS world.

Selfishly, what are the odds that there's an iPhone version in the works? I'm due for a new phone sometime soon, and I imagine you'd get decent performance out of an A12 chip.

_________________
Occupy Babel!

Thanks for developing this. I've also purchased it - and like it a lot.

what I read here smells of awesome!

Guess I need to save to replace my olden iPad 2

_________________
“The only difference between me and a madman is that I’m not mad.” — Salvador Dali ★ Play a slooooow correspondence game with me on OGS?

Kaliwan san,

Thank you for your suggestions!
I had to have a principle to keep it simple when developing this first version to release it early.
I may add setting modal and consider to add silent mode in the next major version,
The functionality for files is also same.
When you add some functionality, you are alway bothered by where it places^^;

Thank you!

casch san,

Thank you for using my app and repeating a bug!
I will investigate KGS SGFs.
I am sorry for inconvenience.

The fill colors and border colors are compatible with Lizzie.
The fill-blue means a primary candidate with most visits in the meaning of Monte-Carlo Tree Search.
The read-border means that the primary candidate does not have highest winrate.
The blue-border means that it has the highest winrate though not the primary candidate.
The gradations of fill colors means a degree of visits.

I hope that you will enjoy my app more.

Grünauer san,

Oh, it must be a bug. I am so sorry.
I will try to reproduce it on my device.
Thank you very much, and sorry for inconvenience.

hyperpape san,

Thank you for your interest in my app!

The app itself works well on iPhone, but 19x19 board is too big for iPhone at present.
So I gave up to release the app as universal one.
I will devise user interface for users to play more easily.
Please wait for a while.

dust san,

Thank you for purchasing my app!
Your message encourage me very much.
Thank you!

Bonobo san,

Thank you for your interest in my app!
I am using an iPad1 and an iPad3 and added a new iPad Pro in the time.
I am so satisfied with the new iPad Pro by my app^^
I hope that next iPad will have A12X Bionic and will be more affordable.

Thanks for your reply. So for an analysis, I have to look for the blue circle (best move) and the upper numbers (winning probability)?

casch san,

Yes, you are right.
The fill-blue move is the one which the AI will play at that moment.
The upper number is a winrate of that move and the lower number is a visits count by Monte-Carlo Tree Search.

When you see blue and red border cirles, it means that the AI is not confident of the fill-blue move yet.
If you give the AI time enough, blue and red border cirles will disappeaer and the AI is confident of the recommended move at this moment.

Sorry for no explanation in the app and I hope that this will be helpful for your Go life^^

Thank you, that was helfull.

Maybe later you can add suport for handicap stones. They are missing when importing sgf.