Would like to test this, too.

PM sent.

hi,where can i download this software?

I don't want to compare it to PCP, I'd rather tell about what my software does well / strives to do well

Xypto is still in beta and maybe that later on people will want to make a thread called "Xypto vs xxx vs yyy" and I'd let the actual user participate. In such a thread, I'd just participate to correct inaccuracies said about my software

I've already mentionned it previously but I think that fast import speed, accuracy, low DB size, graphically configurable HUD that supports unlimited undo/redo (and range-color coding), replayer showing correct calls odds and equities, ease of use (it's really a one-click install), a no-nonsense UI (I hope so and had already someone mentionning here that the UI looked good), etc. are all nice features not necessarily present elsewhere.

And that candlestick/stock-market-like graph for the sessions looks darn good (the idea is not mine that said, it was suggested to me by another programmer who had done it for a personal project)

I think that performances are really good too: I've been able to play nearly 30 tables (in cascading mode) on my old Mac Mini / 1 GB or ram. The limit was me (my unability to handle so many tables at once).


The beta I'm shipping now is unintentionally 'crippled' at about 700 K hands. I'm not focusing development right now on supporting millions and millions of hands because I really need to know first if it installs fine, launches fines, shows the HUD fine, etc.

But honestly I developed something different than what is out there: on a two years old Core 2 Duo / 2 GB of ram, 1 million hand histories are parsed in 8 minutes (including EV calcs) and occupy less than 100 MB in my own DB format. I made tests with up to 1.5 m hands, including 700 K hands for the hero. No problem whatsoever.

So it shall be able to support a huge number of hands.

Also importing lots of hands, managing the DB(s), etc. it all shall be eased due to the speed of the importer/overall engine and due to the compact size of my DB. Say you have a 3 millions hands DB, it shall occupy around 300 MB. It shall not be that hard to back up / transmit from your desktop to your laptop / etc. Now say your DB becomes corrupted (I'll take precaution against that, but it happens, hard disk crash happen too). Well, re-importing 3 millions hands from your hand histories (if you still have them), if you have a Core 2 Duo, should take 25 minutes (probably less than that for people running quad-core). This shall be compelling for people having big DBs: they won't see a progress bar showing an ETA of "x days".

So to answer your question:

- current private beta support around 700K full-ring hands (a bit more than that for short-handed for they take less place) but I advice against running this limit for I didn't put any safeguard in place, so the importer would be stuck on exceptions, trying to put a new hand, giving an error, and trying again with the next hand (I'll fix this soon), which would slow everything to a crawl (I'll fix this real soon actually).

- I've tested with 7 digits hands DB and encountered no particular issues.

- public beta / 1.0 shall support around 1 million hands.

I plan to add later on support for millions and millions of hands, so that people like WCG|Rider and Hodger05 can import all the hands they ever played

Once again as of now I'm looking for people to try the beta and tell me if it works, on which config it works, if the HUD correctly shows up etc.

If you have a Mac you can PM me, I'll send you a d/l link and you'll be able to see that even if it's just a beta, it's already very capable,

Bernard, aka "TacticalCoder"

I can't sent you a PM and I don't know if it's allowed to publish a support email publicly in this 2+2 forum.

I had to post a few messages to be able to PM, it only took a few days.

Don't hesitate to PM as soon as you can PM and if a mod confirms it's ok to publish an email in here I'll do (but I'm cautious about emails because most of them ended up in people's spam folder, which is an issue)

And I'd love to get a programmer's point of view on Xypto!

It's not publicly available yet for it's a private beta, only for the 2+2 community members.

The only way to d/l is to contact me by PM. You're a new user so you'll have to wait a few days/post a few messages, before being able to PM.

Once several 2+2er confirm that the beta is working good enough for them and once I've fixed a few things here and left (including the HUD configuration tool, which I don't like as of now) I'll put my website up and the public beta shall be downloadable by anyone.

But I'm not there yet.

Interesting...

1) Does Xypto reimport all the hands every time it starts up or do you keep the db on disk? If it's on disk, is it a home grown db or some flavor of a SQL db? If it's home grown, can users poke it to construct new data items and run custom reports?

2) Depending on where and when it's accessed, the Pokerstove-like functionality may be against the ToS for most sites.

3) How long before Tourney support is incorporated?

Thanks

First reports came by email by beta testers telling that on the first launch everything was working smoothly, that their HUD launched on Full Tilt / PokerStars on the first try, that they've been able to create their own HUD layouts, etc.

The DB is kept on disk, but not in the beta 0.81 and 0.82 that I shipped for I'm still fine-tuning my schema. So as of now beta testers have to reimport their hands at each new run. But as I already wrote importing is very fast: re-importing 500K hands on a Core 2 Duo should take 4 minutes.

The DB shall be saved on disk before 1.0.

It's not a SQL DB but a very compact Object Oriented DB which I don't want to discuss too much because I'm pretty sure that being able to store 1 million hands in less than 100 MB (including all-in ev adjusted computations) is going to give me an edge over quite a few products

All I can tell is I'm really storing the hands: I'm not just keeping the stats. You could, for example, replay all the hands in your DB.

I construct my own queries using filters. For example to display the hero's HUD stats corresponding only to the last x deals played on each table. At one point it is planned to let user construct their own queries using filters to create their own stats/reports but it's going to be from within Xypto itself and it's not going to be in 1.0.


I gave the stove example to show how fast my equity analyzer was, which has an immediate consequence on the import speed (1 million hands importing in 8 minutes and occupying less than 100 MB includes EV computations). Once again that's something I don't want to discuss in details but it has nothing to do with anything I've read on the forums/papers. In a lot of cases, like say a preflop 3-ways "22+ vs ATo+,ATs+ vs 54+", I'm computing an enumerate-all more than one order of magnitude faster than PokerStove (it takes less than 2 seconds). And the results are, in that case, identical to PokerStove.

I shall never do anything against a site's EULA. If I add a Pokerstove-like tab, it's not going to give an edge anymore than PokerStove gives an edge. It won't analyze a hand currently being played nor fill range automatically. And it certainly shall not give hints as to what would be a correct in game action not display real-time percentage in the HUD based on a hand currently being played.

I'm not reading, not screenshotting, not OCRing the screen and shall never analyze a hand currently being played.

My own forum (only available to beta testers) states in a sticky that I'm not open to any suggestion that would infringe on a site's EULA.

My tracker & HUD will be legit and will be allowed on all the sites where legit trackers & HUDs are allowed.

Support is already incorporated in the engine.

So tourney/SnG/DoNs hands are already imported: my software understands the 'ante' concept, for example.

The HUD already shows up for tourneys (MTT and STT / SnG / DoN).

But as I'm not much of a tourney player I'd need some input as to which kind of reports I should create / which stats the HUD should show and how they should be filtered / etc.

What is missing is parsing the tourney summary files and reporting. I'd like some inputs as to how I should approach that and why not a few complete tourney hands/summaries

Can this program import datamined hands (handhq.com or other hh datamining companies) ?

I guess it's exactly the same as with HM and PT: the importer simply import text files hands (and it does so even faster than what I thought).

Note that importing datamined hands or DB exchanged with friends may get you banned from most sites.

One of my goal is to be an authorized software and I already told beta-testers that I won't be taking into account suggestions that would violate a pokersite's end-user licence agreement.

My tracker is going to be 100% legit like, say, PT and HM.

If people misuse Xypto Poker and violate a site's EULA, it's going to be exactly the same as if they were misusing PT or HM: they'll risk having their fund confiscated but that is not my problem...

Now on the importer being even faster than I thought: on my Core 2 Duo I used to reach 2000+ hands/sec when importing lots of $25NL full-ring hands.

A beta-tester who shall recognize himself (thanks!) sent me a lot of of 6-max hands played at a higher limit. These hands are on average much smaller than full-ring hand and, over a small 550K hands sample, I reached 3024 hands/sec

3 minutes to import 550 K 6-max hands, including correct All-in EV adjusted computations

Any reports from people using it yet ? Poast Up.

Wow this sounds very fast indeed. Looking forward to posts from some beta testers to confirm such claims of speed

Don't have a mac myself as yet, maybe one day...

You say hands are being imported very fast and the DB size for 1M hands will be 100mb. Have you done testing on how fast the import speed is when you have like say 5M hands in the DB already?

1 million full ring hands at micro-stakes shall occupy around 100MB on disk.

I'll make a test with 1 million hand at 6-max/higher stakes and see how many MB it takes and report it here for you. It shall be less than what full-ring hand needs.


I've done test when DB was 1.5m hands, importing 500K additional hands and import speed didn't drop.

The technology used has nothing to do with any tracker out there, so it simply doesn't work the same way. For example when I'm importing a tiny 50K 6-max hands on my test machine, I get 2 600 hands/sec. If I Import 500K (ten times more) 6-max hands, I reach above 3 000 hands/sec and this doesn't depend on the number of hands already in the DB. The importer re-wires itself so that most oftenly used code paths are reached first (so there are less caches miss in the various CPU[s] cache), which is why the more hand you throw at it, the faster it goes. It's a really a cool part the software

That said I shall not be supporting DBs made of 50 millions for the first public beta, nor for 1.0.

1.0 shall support anywhere from 1 million to 3 million hands, I haven't exactly decided yet. It shall be at least one million hand and this should be way enough for most players (not for all players, but for most players) and for 1.0 I'm targetting "most players", not yet "all players"

The reason for this is that I haven't settled yet on my 'final' DB schema. Actually at this point I'm not even bothering to persist the DB on disk yet: beta testers as of now simply reimport their hands whey they launch the program (which is not an issue seen the speed of the importer).

But in any case, size occupied and import speed shall not be modified by the number of hands already in the DB.

Queries are all cached and partial results re-used, so in addition to having a very small DB, queries are lightning fast too.

The DB shall be persisted on disk soon (and use the size I mentioned), but I'm still in beta and as of now beta testers are (fastly) re-importing their hands at each run...

So you are basically importing the hands into your Ram now? When the DB is saved to the harddrive, there will definately be a bottleneck. What about retrieving the stats then in the hud or tracker itself? That can't all be as fast as your "magic" import speeds. It's just like re-inventing the wheel.

Sorry for being that skeptical

Definitely not

No problem about being skeptical...

I won't tell how I did all my 'magic' but I can address your issue...

Import speed shall not be affected by saving to disk (I already did it, it's just not activated in the beta), here's the technical reason:

Importing lots of hands and computing AIEV is CPU-bound (it really makes your CPUs work a lot)... Reading the hands from the hard disk is amazingly fast compared to what the CPU needs to do. Saving the DB to the hard disk is "I/O", not "CPU". I can have another thread (that's technical but I'm explaining it anyway) doing the I/O write to disk while the CPU is "crunching" the hands and doing its magic.

Here's a concrete example that I just ran on an (old) MacBook Pro which is only a Core Duo (not a Core 2 Duo):

- 815 K hands 5-max Ongame (yup, I added Ongame support for two people asked it in this thread)
- time to import : 6 minutes 21 seconds
- hands / sec : 2139 (would be faster on a Core 2 Duo or Quad Core for the importer/AIEV stuff is all correctly multi-threaded)
- DB Size : 62 MB (would be 74 MB for 1 million 5-max Ongame hands)

(exact same 815 K hands on my older Mac Mini / 1GB of ram took 8 min 40 s, running at 'only' 1560 hands/s or so).

A hard disk can easily do a write throughput of 40 MB/s. Let say we'd have a real slow harddisk that does only 15 MB/s. Saving the DB would take 4 seconds... That's the advantage of being compact.

So even if I was only waiting for the end of the import before saving, my 6 min 21 seconds would becom 6 min 25 seconds. 4 additional seconds, a 0.01% additional import time.

Now I'm not going, in this thread (nor on my website), to explain how I did everything I did. Some of the stuff is really advanced and I'm not sharing it.

The software works: using my own dev version I've mass-multitabled PokerStars on that same Mac Mini / 1 GB or ram with a 7 digits hands DB and everything worked flawlessly. That Mac Mini is hardly 'top of the line' and several beta-testers have way more powerful configs...

Anyone from 2+2 that owns a Mac can PM me to help beta-test it. Eventually a public beta shall come out (as of now it's a private beta only for 2+2 members) and then a 1.0.

Skeptical people shall then be able to d/l it and see if for themselves if they like it

But I'm not going to explain how I designed that application (and why I designed it that way) nor how I implemented the 'magic'.

It works, this thread is not a level and the software is for real

Once again, 1.0 may not fit people wanting to store 50 millions hands DB yet but for 1.0 I'm targetting "most players", not "all players".

Of course, I dont want you to share your secrets but what about this part of my post: What about retrieving the stats then in the hud or tracker itself?

It's not slower than retrieving the stats in the other trackers AFAICT... A SQL DB engine does nothing magical that can't be reproduced by another DB

I've been mass-multitabling (nearly 30 tables) on my Mac Mini 1 GB of ram with more than 1m hands in the DB and everything was responsive, HUD updated correctly, etc.

SQL DBs are not the only types of DBs. My DB may even be an RDB without being SQL (SQL DBs are RDBs). Or it may be a column-based DB, or a correlation DB. But I'm not telling

Beta is out, people can already load 1 million 6-max hands (a little less full ring) and try for themselves to see how the tracker and the HUD works

I got the beta yesterday and all I can say is: WOW!
Some facts I got out of it:
- Installing = 1click, no additional postgre, nothing. click, start, be happy. worked perfectly.
- I'm importing at 3.3k hands/sec (yes, I really do) with 2.4ghz dualcore and 4gb ram macbook pro.
- the interface of it looks already quite nice and is really responsive and fast.
- there are actually some ideas (its still BETA so whatever) that might have a lot of potential for all the grinders out there (one example: nice autoimporting overview with all kinds of stuff (like the HMhud window) but within the program as a tab - could be used (atm it is just an overview) to customize display of several stats without looking at a session report (time played, hands played, (dollars won), MT ratio, hands/hr...i think this could be great for a lot of grinders including myself
- there are still bugs in it and its nowhere near COMPLETE but you can already play and grind with it

the most impressive thing is the speed of the whole program though. you'll never think of such hyperspeeds when you use pt3 or hm. all the stuff you can do is sooooooo fast.
also the developer really tries to improve and bugfix everything ASAP, i think he's really putting in a lot of time in it and thus i smell that finally there is a (very) good OSXtracker on its way which MIGHT actually crush HM and PT3 because it's so damn fast and all the innovating stuff can flow in it if you test it too.

VERY interested in beta testing this software. Have sent PM and email.

If you are still looking for beta testers I am more than interested.

Have tried it and it IS what he says it is. It's very fast to import hands. Also, very well laid out. It is still a beta, so you do have to realize that it's not going to be very close to the final product. The stats and features are there. You know, the ones that us Mac users want like the rest of the PT3 and HEM users have. The HUD is basic at the moment, but I can tell it should be on par with HEM in the future.

This guy is putting his work out there for Mac users to evaluate and help improve the product. And so far, has been very quick with responses and answers to any issues that may pop up. Give him a shout and try it, you may be surprised. Just remember not to expect a candy coated final product, YET. This is going to be outstanding software as long as we can support him and be objective.

BTW, I am a big fan and user of PokerCopilot. I have used and enjoy (and still do) his product as well. Very well done for a one-man software development company.

Oh, and if you think PT is going to come out with a Mac version anytime soon, you're just kidding yourself. All they have been doing for the last year and a half or more is saying 'should be available by spring/summer/fall/someday'.

yeah PT is driving me crazy!! i <3 there product (i spent alot of time figuring out how to get the most out of PT2) and would by it in a sec, but if they seriously cant get there **** together by the end of the year, i'll either go with this or copilot or get a PC/run windows on boot camp and use HEM. Its really pissing me off, but im not putting in too much volume anyway so im not too pissed about it atm.

on to Xypto - i had issues getting the HUD to work on 6max tables tonight. it popped up with an error message but im lazy so im not gonna try and replicate it.

Ah... The current beta isn't supposed to have error message in the HUD that pops up

It's either "HUD: Waiting for a hand on table 'xxx'" or "HUD: Play money tables not supported" or "HUD: Licence problem, beta too old", which may happen if you're running beta 0.81 or 0.82 and not 0.83.


Getting the HUD to work the first time and then getting used to its quirks may take some time.

Instead of doing a FAQ and a checklist, I decided to incorporate a checklist directly in the software:



But I think I'll change the terminology a bit and try to make it easier to troubleshoot.

For the HUD to display stats, there must be an entry in '3' (one per table). For an entry to be in '3', the table should both be found on screen ('4') and from a recently parsed/auto-imported hand-history file.

If the "HUD: Waiting for a deal on table ..." shows up ('6') and if the table name is correct then things start to look good...

Apparently Dazbog got used to the HUD's quirks very quickly: he already managed to get the HUD working using a workaround on a dual-monitor setup, which isn't supported yet in the current beta

It really shouldn't be hard and if it's hard, then I have to simplify the procedure

I'll also change the "HUD Configuration" tab name to "HUD Layout Editor".

Some people never see the stats showing up, but I think it's due to some silent internal bug. I'm working on it

sent pm about testing beta

Same here, I'm gonna leave my home for a few days and will only be able to play on my mac book, I would really like to try this instead of reinstalling vmware.