As of November 2021, the APT Forum is closed to new posts. Like with many online forums, usage has decreased in recent years. All previous posts are still available.
ROI Leaderboard on Final Table Trainer Math Issue
Can someone take a look at the ROI stats on the Final Table Trainer for the 30 and 365 day tabs? How exactly would it be possible to achieve ROI's of 20,000,000%? Theoretical maximum should be 85,000% at hardest level. I'm thinking there is a bug on the 30 and 365 day calculations as the 180-day with decay numbers are in line with theoretical maximums. The issue is the same on both the hardest level and all level tabs, where 180-day numbers are fine but 30 and 365 day numbers are out of whack.
ADDED: OK, I looked at it and I think the problem is that the software is dividing cumulative earnings by a single buy-in rather than the sum of all the buy-ins.
Sign In to comment.
Welcome
It looks like you're new here. If you want to get involved, click one of these buttons!
Comments
Those are some awesome ROIs lol. Thanks for the info, we'll check it out.
Thanks. If I ever saw numbers like that for real, I'd quit playing poker, quit my job and just buy pieces of their action
It is fixed now. I think you will like it a lot more warlock, especially since you're at the top of the leaderboard now
If that's the case, then you may have more bugs in that thing than either of us knew about
Added: no wait. I moved up from 16 to 11. Nice changes you made there.
LOL - has to be rigged. Now if I could just figure out how to do it for the 9-max game. You still have me by about 63BB/100 there. I cut my losses to ~11BB/100 but that's a big old hole to try and dig out of. I'll be happy to just get positive this month on full ring. I've run colder than a snowman's jockstrap so far.
The FT trainer ROI still seems incorrect.
Yesterday, I played 5 times with a 49k ROI and yet my leaderboard ROI went down from 31k to 30k for both the 30 & 180 boards.
All my ROIs from last 24 to lifetime are above that.
Can you check it again?
Hmmmm, I'll check it out tonight...maybe there is still a problem.
So, it is right now, I computed it by hand just to check. I think it went down either because of the changes we made yesterday (that took effect at midnight), or maybe you are forgetting about 5 FTTs you played on July 29th, that were at the $100 buyin "Vegas Daily Tournament" level, but still on the hardest skill setting, so they count.
Also, we compute ROI individually for each tournament, and then average them all out. That is different from a traditional view of ROI (total net winnings / total cost). We do it that way so FTTs played at the hardest skill setting, regardless the buyin, affect your ROI equally.
Steve
Interesting way to calculate ROI. I never would have known had you not stated it here. Would it be possible to add explanations and formulas for how results are calculated for all the leaderboards and personal stats pages? Maybe its just me but I'm wonky and I like looking through the details so I can construct the best strategies for the structures. You should see how I go through tournament structures and payout tables before deciding which to enter. I've always believed that attention format and structure are the most underappreciated aspects of being profitable at this game.
Sorry for all the questions and requests. Consider it repayment for the angst your bot is causing me lately
Dang thing plays some good heads-up poker.
warlock my friend and helper and so brilliant!
Mr Blay if he giving you problems let me know ok?
Can you tell me which personal board you're referring to?
And yes, those will disappear after 30 days.
Not a huge thing. The difference is peculiar though.
@SteveBlay and @AllenBlay - hate to do this to you but I think there is another math bug on the leaderboards (Weekly Challenges this time). I took the latest challenge and ended with 11,688 chips (it was a good week). When I go to my 30-day stats on the leaderboard page for the previous 30 days, it shows my average is about 900 chips. That isn't possible, even if I scored 0's in the past 3-4 challenges. If there were 5 challenges in the previous 30 days (as a maximum) and I had 4 0's and 1 11,688, my average would be over 2,300, not 900.
Looking at my saved sessions page, I see I have 4 challenges taken in the past 30 days. Ending chips were 2,831, 2,733, 0 and 11,688. That averages out to 4,313. Slightly more than a rounding error away from 900
I didn't go through the data on the 180 or 365 day pages as that would be too time consuming for me and I doubt I even have all the data anymore as I used to just delete sessions that didn't provide any meaningful data for me to use. Still, given what seems to be a problem with the 30-day results I wouldn't assume the other pages are correct at this point.
I do enjoy this site and just recently became interested in the leaderboard thing because of @highfive. I hadn't looked at the individual boards until just last week and already I've noticed at least 1 glaring error and maybe 2 if I'm not mistaken here. I like competition as much or more as the next guy and I think the leaderboards could be a motivator for players here. However, if the stats are incorrect then the utility is lost. I'm sure I'm not the only one who would feel better if the site had fewer errors. Finding them leads me to question what other errors there might be and whether they may occur in critical components of the training functions. Maybe take a look through all the leaderboards and make sure the data is being pulled correctly and then crunched properly?
@1warlock
@highfive - gone through my stats and there are several errors I've spotted. Don't know if the data is bad or if its pulling from the correct places or if the formulas are off. You know where I stand and it may be time for me to move on before I had planned to.
To which session are you referring highfive, and on which screen? I'll look into it. Also will look at the weekly challenges warlock. My web programmer is on vacation but I should have an answer sometime next week.
I would look @1warlock's weekly challenge. It shows he surpassed 11k but LB average is 900+.) There's no way unless the LB doesnt update immediately.
Also he has cashed both MTTs but has -ROI. Again. No way.
I am unconcerned with mine. The session with 675 hands if you really want to look as it.
Prioritize Warlock's situation.
@highfive - I don't think its a matter of prioritizing anyone. To me its more a matter of making sure the site is returning the correct information to the correct places. I think both you and I are seeing some numbers that don't make sense and want them addressed as part of an overall quality control review.
Deleted due to further information received
ok guys, I'm on the case...but I need a week or so until my programmer is back and we can sit down to look at these together. It would be favor to me if no one post any more bug reports to this thread until I get a chance to look into these. (And, there will be a logical answer for where those numbers are coming from.)
@SteveBlay - no worries. Comments amended to reflect your response.
- Only random hand sessions without any peeking at opponent cards are included in the leaderboard session. Everything you do goes into your reports. This makes sense - if you are looking at reports for your own benefit, it should show everything unless you filter it out. But if you are going to compare apples to apples and rank people, it needs to be the same things. So a common question we always get is why reports aren't equal to leaderboards. The solution to that is more detail in the help file - but we still get questions all the time even after putting the detail in there. It is sort of the nature of flexibility - we let the reports be flexible, but not the leaderboards. Deleted sessions are another huge source of confusion because they don't show in reports, but obviously count in leaderboards (more on that in a bit).
- The weekly challenges are even worse for trying to explain what is included for the leaderboards. If we really want this to be comparable, all users have to have played the same challenge types. So for a 30 day leaderboard, no matter what we include there will be someone confused if we include the same challenges for everyone. For example, with the current week AA challenge, lots of people hadn't played it yet. Would it be fair to include it for @1warlock before everyone had played it? No, he would have one more challenge than everyone and it would be a big win challenge since it is AA. So we don't include challenges that haven't completed (which I think is the right way to do it). But on the early side, some challenges started more than 30 days ago, but some people played in it within 30 days. In 1warlock's case, the one on the front end was played by him within 30 days - but it actually started more than 30 days ago and 1/2 the people who played it didn't play it within the past 30 days. So for fairness, challenges that began more than 30 days ago aren't included. Obviously the effect of this is much, much smaller on longer intervals like 180 days. But for a 30 day period, it can lead to situations like the one posted where all a user does is see what they personally have played in the 30 day period, and the numbers don't seem to make sense. In reality, for leaderboard purposes, they are actually correct. 1warlock had three challenges that were completely contained in the 30 day period: 0, 2733, 0. 2733/3=911, which is exactly what it was showing for average chips. And 2733 is what is showing for total chips. The one he listed with 2800+ chips started more than 30 days ago, so it isn't included. There also was another one with 0 that wasn't listed in the post - it was deleted in his sessions so he didn't see it, but even sessions a user deletes are counted in the leaderboards for fairness (yet another possible source of confusion).
So the bottom line is that the calculations are correct, just the nature of what is included is incredibly difficult to explain while keeping it a fair apples to apples comparison.
@1warlock was correct about the FTT leaderboard. So I'm not sure where this confidence in the acurracy originates. I also find it suspicious that after the formula was altered warlock is on top. (Joke)
Steve had requested our patience. Warlock agreed. I complied. I have a few more if this thread is to resume. I think an overall quality control review is in order.
My mention of the 6 max session was about that session only not with respect to the leaderboard. I've likely fallen out of that by my own poor play in 9 max prior to the 6 max session mentioned.
The FTT board was wrong for 9-10 months so a natural question is, " what else is there?"
@AllenBlay - thanks for the explanation. That makes sense, especially on the weekly leaderboard. We thought it could be something along those lines but the leaderboard states up top that it updates every day. Same thing under the FAQ section where reasons are given for why someone is or is not showing up on a leaderboard (hasn't yet updated at midnight EST). Maybe change the caption on that board to indicate a once weekly update?
In my particular case, I did look at what looked like the relevant date range but I obviously didn't select the correct ones. Up until very recently I deleted sessions that were either too short to give me relevant info or that I had to abandon at some point (like having to work or attend to something else) or any number of other reasons. Only in the last week or so did the leaderboard thing come to my attention and interest. So, if missing a zero and skewing the dates would produce the posted answer, problem solved. The data is good but my understanding of where it was being pulled from was not. I can certainly accept that and offer my apologies for any stress I may have caused. After the FTT issue, I think it was reasonable to question other results that didn't seem to make sense.
As to the issue of "fairness", well I'm kind of old-school when it comes to this. If everyone is given a clear set of rules and standards, then that's as fair as it can get, especially in poker. I've seen debates on other sites about how this leaderboard favors volume of play or the other one does something else. People complain about all sorts of things. My view is that the host gets to make the rules and anyone interested in that particular competition can accept them and try or not. Yes, the players don't play duplicate hands in the Weekly challenges so that can skew results for that category. However, we also don't play identical hands for 10 of the other 11 categories. Some people will run good for a month and some run horribly. That's just life. My point is that a host can set up whatever contest they wish. They don't have to offer any at all. Is that more fair? Probably but its also less interesting.
I do have a few more seemingly anomalous pieces of data that I want to examine. I am not ready to say that all the data is correct at this point. After all, we did see 1 massive error before. In the particular case I brought up, it appears that I was incorrect but also that the site was not very precise or clear about how those results were being tabulated. The daily update notation is incorrect.
I'm glad the site looked into this and that I was mistaken. I'd rather be wrong in this case.
@1warlock - I enjoy the dialogue and looking into the data. I think that's why Steve and I are both so active in this and respond to everything ourselves. We are doing a pretty comprehensive look into the numbers in the leaderboards etc. just to check up on things. We were pretty surprised to find the other issue because we launched those leaderboards almost two years ago. Usually we find bugs pretty quickly because either we notice something bizarre or someone else does. With as many calculations as there are on the site, things do pop up, but usually a lot sooner than two years down the line. The fact that this one was so completely mathematically impossible, yet hadn't been noticed in over two years by anyone made us realize how little anyone really pays attention to the leaderboards or the numbers on them, which is a big part of the reason we're spending a little time on this just to make sure there isn't anything else weird floating around.
So, about the weekly challenge leaderboard. It isn't possible to show the current active challenge in the leaderboard because of the "Most Top 10 Finishes" ranking. And obviously we don't know that until the challenge ends. So to avoid the two "views" using different challenges, all of the weekly challenge leaderboards don't include the current week's challenge. I added a message at the top to clarify that: "Please note the current active challenge is NOT included in these rankings."
There was a problem though. Sometimes a challenge you played, say, 27 days ago, was not appearing in your Last 30 Days results, because the actual challenge itself (for that week) always starts Sunday at 12:00AM, and it was using that date. I've fixed that now, so any challenge you took in the Last 30 days (or any other time period) should show up correctly (again, assuming it isn't the current week's challenge)
Highfive, I've looked back at that session with 675 hands and don't see a problem. I added up all the hands played and the net total is $-64, which over 675 hands comes out to 1.9 BB/100 lost.
Finally, I think I already addressed the Final Table Trainer leaderboard, and fixed the problem there. Thanks for your patience.
I went -2ish buyins and rallied a few buyins to the positive. I did the math at the time. It should have been +30ish bb/100 over that many hands.
We will drop it. I was having connectivity issues then so i dont think the data transfered correctly.
One last question remains @SteveBlay. Why is my FTT leaderboard 30k yet my 30 day ROI and lifetime ROI 39k?
Thanks @SteveBlay for the update. The weekly challenge sounds like a challenge to work with
This helps explain why I was looking at the wrong data in my saved sessions reports. I was going by actual date, not starting date of the particular challenge. Hey, so long as everyone knows what's included and what isn't, its all good. As a follow-up question: The leaderboard for the Weekly still indicates a daily update. I thought you weren't updating it until the challenge was over. Where did you come down on that? If still the same, then maybe change the daily update note to indicate it will only update after the current challenge is over?
Best and thanks again. I don't think I've found any more results that seem off but if I do, I'll let you know.