Monthly Archives: November 2014

Hybrid update

I’ve been testing the latest updates solidly since the last post and it all works perfectly except I identified a situation where if you don’t have enough spins for analysis, it will be too easy for an inexperienced player to have one of the settings adjusted incorrectly. It’s no problem if you have enough data, and it is a rare situation, but frequent enough to need addressing.
 
There’s an easy solution that’s being applied and I expect to release the next version before the weekend, or latest on Monday. I know this particular release has taken more time than expected but it includes a lot of extra parts and I’ve had to thoroughly test them, and I’m sure players can appreciate I wont release anything that is less than perfect. The alternative is I don’t test properly and release it anyway, and it can lead to losses. So I appreciate everyone’s patience to ensure that profiting is as foolproof as possible.
 

Hybrid mod update

Target/dynamic is fully complete, tested and working perfectly. The accuracy improvements are even better than expected. The mod specifically for players aiming the camera at monitors/screens is tested and also working perfectly.
 
But before I provide the next release, I want two small mods to be done (otherwise I’d just be issuing another release right after):
 
1. Especially for slower hardware, the recorded video runs at a slightly slower speed than real-time, so setting up will not be as accurate is it should be. The difference that would otherwise cause inaccuracy is automatically tuned out during actual play, but it takes a few more spins than it should.
 
2. Minor interface changes so it is cleaner and simpler to understand.
 
All the major development is done, so now we’re mostly cleaning up. Still though there are parts to add later but nothing is overly important. It’s more about making winning as foolproof as possible. The whole idea of the hybrid from the start was not just best possible accuracy, but also for anyone to be capable of winning with barely any instruction. The target/dynamic mod was a big part of this, because now you don’t need to worry about peaks being ok, or if rotor speeds are ok. The software will just announce “risk” if the ball isn’t likely to bounce in the most predictable way and with a clear edge (based on rotor speed and where the ball is likely to fall from the ball track).
 
I’ll never say there can’t be any further improvement and will always look for ways to make it better. Although clearly with current modern wheels and conditions, it doesn’t need to be any better considering you need to deliberately win less to avoid detection.

General schedule update

For the past week I’ve been busy for the launch of my new website www.landsharing.org – it’s basically a website for people to share land and grow food, instead of letting land go to waste. It also connects people who want to live a debt free life, with the same kind of principles explained at www.yourway.org.au. In the coming weeks, the landsharing.org website will be officially launched, although it’s now fully operational if you would like to use it. It’s completely free forever. At the start of next year, related project will be taking more of my time. But for now the hard work is done and I’ll be up to date on emails by the end of tomorrow.
 
Also I have received an updated version of the Hybrid but will need to test it tomorrow. The programmer has said the minor issues with target/dynamic are now fixed, but it will take about 2 hours of testing to confirm.

Hybrid Dynamic Mod

Testing revealed a few minor issues, some of which are now fixed but there are still a few additional parts to fix. They don’t reduce accuracy, but they make the software call “risk” when it shouldn’t. Issues are being worked on. Besides this the testing results show exceptional improvements in accuracy.

Hybrid wireless camera mod

One of the upcoming Hybrid mods will allow the use of the wireless camera. It already can do this with additional equipment, but this mod will remove the need for any additional equipment. This is so the only thing the user will have on them is a mobile phone. The main processor of the Hybrid can be far away from the table.
 
The video transmissions are properly encrypted and virtually impossible to detect. The wireless camera also uses it’s own computer to deal with erratic camera movement, so you will be able to “jump around” more freely even if using the hybrid version that has restrictions on hardware speed. You can already jump around quite freely, but within reasonable limits. This mod will increase the limits. So basically we are using multiple small electronic devices to share the processing load. As more powerful “small” devices become available, we can start using the full algorithms.
 
This mod is a bit of overkill because the main unit is already small enough to hide while at the table, but it will be useful for situations where casinos use metal detectors at the door. This is very rare although still having this mod provides players with greater flexibility. Also I want to see the Hybrid do everything that can be done, even if just from a hobby perspective.
 
There are two parts to this mod, and coding has started on the first part today. The programmer for the main processor will work on his part after a few other mods are complete. It should all be available within a few weeks.

Hybrid’s dynamic target mod complete

This mod is now complete. From testing it very accurately targets specific spins where the ball bounces most predictably. For most wheels, this will be spins where the ball hits a diamond and falls straight down. You also have the option of how accurate you want the predictions to be. In average conditions, this mod provides an accuracy increase of between 30-500%. This means instead of getting a 10% edge on a very difficult wheel, you could get a 60% edge with this mod.
 
These projections are not at all exaggerations. In fact they are actually conservative because some types of ball falls are very easily predictable, and others are barely predictable at all. This mod allows you to isolate only to spins where the ball is most likely to bounce very predictably. Of course it wont be correct all of the time, but it is correct enough of the time for a large accuracy increase.
 
In many cases, this mod is essential to get any edge at all. Such a situation would be if the rotor speed was varied considerably (between 2 – 8s / revolution) and the ball is quite bouncy and dramatically affected by the varying ball bounce of different rotor speeds. It is rare that you will “need” this mod to profit, but almost always it will significantly boost accuracy.
 
The downside of this mod is on some of the spins, the computer will announce “risk” indicating that the predictions are not as accurate as they could be. You can set the approximate proportion of spins that are skipped. Basically the more spins you skip, the higher the accuracy. Putting it into perspective, if you get 1 prediction from 3 spins, then you may increase your edge by about 200% (ie a 10% edge becomes 30%). A larger edge is not everything though, because the fewer spins you bet on, the less you may win “per hour”. But a critical point is if you bet on every spin, then your behavior is too obvious for casino staff. So it is better to bet only on the most accurate predictions. It comes down to what I often say “it’s mostly about what you can win without being detected”.
 
The next mod will be started soon. It will improve accuracy in cases where the camera is aimed at a computer monitor of some kind.

Android roulette computer 30% complete

The Android programmer has been working on development full time, and all components are already about 30% complete. There are server components that must be done, although the programmer that will work on those parts are busy working on other projects. He’s a programmer I’ve been working with on other projects for some time, and he’s very good at what he does. So it’s possible the completed computer may be available by the end of November. It’s mostly a matter of when he finishes other work first. It will take him approximately 2-3 weeks once he begins his part.

JAA updates

The programmer has taken too long to complete the JAA updates, so I’m getting a new programmer for it. The updates are so incredibly close to completion, but he didn’t seem to understand the requirements and kept wasting time. The mods are quite technical but he has had enough time to complete them. I’ll need to select a new programmer after I’m back around 10th November. This doesn’t affect the existing version or users of JAA.

Away for most of week (2nd-10th Nov)

I’ll be away for most of the time between 2nd-10th November. I’ll still have email and Internet access, but responses will be delayed and I wont be taking calls. I’ll still respond to urgent issues and code requests. Everything will be back to normal after the 10th. Of course work from programmers will continue without interruption.