After the listenthrough all in on Jadakissnia to be the runaway winner

PopHub is defo the contender from semi 1 I think but dare I say Lotunia could dark horse?

Jadakissnia definitely have a good shot from the semi 2 lot and I say that without hearing any other songs 😂

Here are the semi predictions from Baleghde this month, again based solely on previous voting patterns and who's in which semi final - song quality is not taken into consideration at all. Semi final 2 was a bit more difficult to calculate because a good few of the entrants have not had the chance to vote for each other before, so I had to weight in some other data in those cases.

Semi Final 1

1 Taahino (115 pts)

2 Danæviia (112 pts)

3 Kathaldazia (98 pts)

4 République d'Áskorza (93 pts)

5 Altyr Muunilinst (90 pts)

6 Fljótavík (82 pts)

7 Terra Avium (74 pts)

8 New Lexico (72 pts)

9 Thelonia (70 pts)

10 Baleghde (68 pts)

11 POPHub (62 pts)

12 Mothertopia (62 pts)

13 Land of DW (62 pts)

14 Why So Sirius (61 pts)

15 Lookylion (56 pts)

16 Summericia (55 pts)

17 Flynnonda (51 pts)

18 Cheiron (49 pts)

19 Lotunia (48 pts)

20 Aelandor (43 pts) (breaks the tie)

21 Aesthetica (43 pts) (tied)

22 Harmonica (40 pts)

23 Pagasa (38 pts)

24 Herbilore (38 pts)

25 A-Konohagakure (22 pts)

26 East Rukahavian (20 pts)

Semi Final 2

1 Jadakissnia (124 pts)

2 Aeroche (111 pts)

3 Skall (104 pts)

4 Persephonia (101 pts)

5 Elëdan (88 pts)

6 Cumulonimbia (88 pts)

7 Kluminican Republic (80 pts)

8 Trifoski (74 pts)

9 Hendinia (72 pts)

10 Deandria (62 pts)

11 Scotland 2 (59 pts)

12 N DNTN (57 pts)

13 Greenfroze (56 pts)

14 Cor Lupus (54 pts)

15 Bronzil (54 pts)

16 FARC (51 pts)

17 Ajanaeda (50 pts)

18 Suedeonia (46 pts)

19 Zanmatony (44 pts)

20 Dobago (43 pts) (breaks the tie)

21 Bøtanikkä (43 pts) (tied)

22 The Spodic Empire (39 pts)

23 Pestolia (36 pts)

24 Sandénia (35 pts)

25 Séyetana (33 pts)

26 Jupiterdonia (20 pts)

Jafet, stop influencing the results outcome!!!

Ya'll that's FAKE NEWS!!!

I would expect East Rukahavian will be a lot closer to the other end of the scoreboard on this occasion. (Also a big no to some of those other DNQ predictions)

I'm here for POPHub to be a win contender but obviously my support for it means it's not winning x

Bøtanikkä apparently DNQing from a purely voting history based prediction is interesting considering they have been one of the most successful countries of the last year! Does not seem like an ideal semi draw for it tbf.

how to receive votes in a consistent enough fashion to qualify on Jafet's algorithm

Even more nervous than I was before with my semi draw after Jafet’s algorithm 😅

I’ve really enjoyed the semi though even though there are countries I don’t usually share votes with. I actually struggled to dwindle my votes down to ten.

1 hour ago, diva thin muffin said:

Here are the semi predictions from Baleghde this month, again based solely on previous voting patterns and who's in which semi final - song quality is not taken into consideration at all. Semi final 2 was a bit more difficult to calculate because a good few of the entrants have not had the chance to vote for each other before, so I had to weight in some other data in those cases.

Semi Final 1

1 Taahino (115 pts)

2 Danæviia (112 pts)

3 Kathaldazia (98 pts)

4 République d'Áskorza (93 pts)

5 Altyr Muunilinst (90 pts)

6 Fljótavík (82 pts)

7 Terra Avium (74 pts)

8 New Lexico (72 pts)

9 Thelonia (70 pts)

10 Baleghde (68 pts)

11 POPHub (62 pts)

12 Mothertopia (62 pts)

13 Land of DW (62 pts)

14 Why So Sirius (61 pts)

15 Lookylion (56 pts)

16 Summericia (55 pts)

17 Flynnonda (51 pts)

18 Cheiron (49 pts)

19 Lotunia (48 pts)

20 Aelandor (43 pts) (breaks the tie)

21 Aesthetica (43 pts) (tied)

22 Harmonica (40 pts)

23 Pagasa (38 pts)

24 Herbilore (38 pts)

25 A-Konohagakure (22 pts)

26 East Rukahavian (20 pts)

Semi Final 2

1 Jadakissnia (124 pts)

2 Aeroche (111 pts)

3 Skall (104 pts)

4 Persephonia (101 pts)

5 Elëdan (88 pts)

6 Cumulonimbia (88 pts)

7 Kluminican Republic (80 pts)

8 Trifoski (74 pts)

9 Hendinia (72 pts)

10 Deandria (62 pts)

11 Scotland 2 (59 pts)

12 N DNTN (57 pts)

13 Greenfroze (56 pts)

14 Cor Lupus (54 pts)

15 Bronzil (54 pts)

16 FARC (51 pts)

17 Ajanaeda (50 pts)

18 Suedeonia (46 pts)

19 Zanmatony (44 pts)

20 Dobago (43 pts) (breaks the tie)

21 Bøtanikkä (43 pts) (tied)

22 The Spodic Empire (39 pts)

23 Pestolia (36 pts)

24 Sandénia (35 pts)

25 Séyetana (33 pts)

26 Jupiterdonia (20 pts)

If I come last with this entry I'm seriously considering quitting 🤣

Ok so my aim for next few months is to get some good results so people don’t put me in the NQ category

I wanna note that I was quite far off last month so pls don't take my "prediction" too seriously

I think East Rukahavian being last is enough to render that meaningless but we love a bit of statistical cat being thrown amongst the pigeons x

11 hours ago, diva thin muffin said:

Here are the semi predictions from Baleghde this month, again based solely on previous voting patterns and who's in which semi final - song quality is not taken into consideration at all. Semi final 2 was a bit more difficult to calculate because a good few of the entrants have not had the chance to vote for each other before, so I had to weight in some other data in those cases.

Semi Final 1

1 Taahino (115 pts)

2 Danæviia (112 pts)

3 Kathaldazia (98 pts)

4 République d'Áskorza (93 pts)

5 Altyr Muunilinst (90 pts)

6 Fljótavík (82 pts)

7 Terra Avium (74 pts)

8 New Lexico (72 pts)

9 Thelonia (70 pts)

10 Baleghde (68 pts)

11 POPHub (62 pts)

12 Mothertopia (62 pts)

13 Land of DW (62 pts)

14 Why So Sirius (61 pts)

15 Lookylion (56 pts)

16 Summericia (55 pts)

17 Flynnonda (51 pts)

18 Cheiron (49 pts)

19 Lotunia (48 pts)

20 Aelandor (43 pts) (breaks the tie)

21 Aesthetica (43 pts) (tied)

22 Harmonica (40 pts)

23 Pagasa (38 pts)

24 Herbilore (38 pts)

25 A-Konohagakure (22 pts)

26 East Rukahavian (20 pts)

Semi Final 2

1 Jadakissnia (124 pts)

2 Aeroche (111 pts)

3 Skall (104 pts)

4 Persephonia (101 pts)

5 Elëdan (88 pts)

6 Cumulonimbia (88 pts)

7 Kluminican Republic (80 pts)

8 Trifoski (74 pts)

9 Hendinia (72 pts)

10 Deandria (62 pts)

11 Scotland 2 (59 pts)

12 N DNTN (57 pts)

13 Greenfroze (56 pts)

14 Cor Lupus (54 pts)

15 Bronzil (54 pts)

16 FARC (51 pts)

17 Ajanaeda (50 pts)

18 Suedeonia (46 pts)

19 Zanmatony (44 pts)

20 Dobago (43 pts) (breaks the tie)

21 Bøtanikkä (43 pts) (tied)

22 The Spodic Empire (39 pts)

23 Pestolia (36 pts)

24 Sandénia (35 pts)

25 Séyetana (33 pts)

26 Jupiterdonia (20 pts)

Yes do this one!

11 hours ago, diva thin muffin said:

I wanna note that I was quite far off last month so pls don't take my "prediction" too seriously

I wonder if injecting a bit of entropy and randomness into it would make it a bit less like a 'past performance simulator', if only because past performance unconscious bias is already something we want to discourage out of predictions. Not that it's going to match anything anyway but would help with simulating a real contest with given semi-makeups where some countries will always (or at least should on average) overperform or underperform any calculation on voting patterns.

Also my suspicions (don't know exactly how you've calculated but this is what I suspect) are the amalgamation of voting stats causes people who frequently receive points from a smaller set of voters are advantaged where as people who get points over several contests from a much wider list of voters get ranked lower on all of them. Which is fine for the second group in a singular semi as they might get high points from a few there to qualify, but then get it from a different few next time, whereas if the first group also qualify with high points from the same list of voters in those two contests, they get much more of an advantage when the stats are amalgamated like this. Does that make sense? Even if on average half of the bigger fans of the first group are separated by semis, on a wider scale, that means that two countries with the same level of points success will always have one doing better than the other. If I'm way off please correct me.

don't get me wrong it is a cool concept to simulate contests and there's potential there but after seeing a second run of it I'm picking up some biases that make it less of a prediction and more a 'YTD sorted by vote concentration and divided by semis'.

@Iz 🌟

I actually already do add a little randomness factor. As for your assumption, I'm afraid I don't understand what exactly you are saying to be honest. For what it's worth, I'm not trying to claim that my prediction is accurate in any way. It didn't even manage to accurately predict my own votes. My top score wasn't even in my predicted top 10 lmao. As for your last comment, I'm not sure what to say but I've already tweaked the algorithm quite a bit between months, and the data I'm using goes a lot further back than just the past year now.

For clarification, it takes into consideration:

• Points given in total vs Max possible points in the same time period

  • In other words, if Abcada has voted in 3 finals and 2 semi finals where Effeghe was competing, but only gave 3 points in total, their score would be 3/(3*18 + 2*12) = 3/(54+24) = 3/78 or 3.85%

• Additionally, I also take into consideration how often country A gives country B any amount of points vs how often it has had the opportunity to do so.

• Finally, I add a little randomness factor to get a little variety.

Hope that answers at least some of your questions :) but again, this is not meant to be accurate. All it does it allow me to make a prediction without my personal bias showing lol. And yes, there are obviously some patterns that will show up, since some people do very consistently vote for the same countries but then again some countries also do very consistently do well in the contest so maybe this pushes people to try to be aware of their own biases when voting.

But if this upsets people I can certainly refrain from posting it again and just keep it to myself.

I love it, keep posting! I think it's very accurate for me anyway I expect this entry is gonna be super borderline.

And I do think form is a very big factor in BJSC. I find that usually when you're on a good run it's easier to qualify, for whatever reason. That's why all the top 10 in EOY keep pulling away and the also rans like myself just keep nabbing the odd qualification here and there. When you're in poor form you tend to overthink entry choices and end up sending bigger risks.

I just noticed no one has actually predicted either semi yet 👀

Who will be the most successful returning nation this month? Deandria? N DNTN? Harmonica? Someone else?

Indeed, let's not influence the current results, my plan exactly.

(I might bring a prediction together after I vote later)

10 minutes ago, diva thin muffin said:

@Iz 🌟

I actually already do add a little randomness factor. As for your assumption, I'm afraid I don't understand what exactly you are saying to be honest. For what it's worth, I'm not trying to claim that my prediction is accurate in any way. It didn't even manage to accurately predict my own votes. My top score wasn't even in my predicted top 10 lmao. As for your last comment, I'm not sure what to say but I've already tweaked the algorithm quite a bit between months, and the data I'm using goes a lot further back than just the past year now.

For clarification, it takes into consideration:

• Points given in total vs Max possible points in the same time period

  • In other words, if Abcada has voted in 3 finals and 2 semi finals where Effeghe was competing, but only gave 3 points in total, their score would be 3/(3*18 + 2*12) = 3/(54+24) = 3/78 or 3.85%

• Additionally, I also take into consideration how often country A gives country B any amount of points vs how often it has had the opportunity to do so.

• Finally, I add a little randomness factor to get a little variety.

Hope that answers at least some of your questions :) but again, this is not meant to be accurate. All it does it allow me to make a prediction without my personal bias showing lol. And yes, there are obviously some patterns that will show up, since some people do very consistently vote for the same countries but then again some countries also do very consistently do well in the contest so maybe this pushes people to try to be aware of their own biases when voting.

But if this upsets people I can certainly refrain from posting it again and just keep it to myself.

Ah, no it's useful to see exactly how it is working. I was noticing similar countries at the top and the bottom in each run and trying to deduce how it might work and therefore assuming that it would output very similar results each time it was ran.

I do think it does then advantage concentrated support based on that, right? To explain my assumption with an example: Two countries each qualify from a given semi with 60 points each, 10 contests in a row. Country A receives 5 12s from 10 other countries, and each contest has 5 of those in their semi. Country B's 5 12s come from 20 different countries, and each contest has 10 of them in their semi, receiving 12 from half of those and 0 from the other half. When it comes time to calculate what every country will give Country A & B, Country A is at the top of 10 lists and receives top marks from however many are in that semi. Country B will be a little way down most of those lists and may receive 5-1 or nothing from their 20 supporters, who don't support them all the time. Expected value for A: 60, expected value for B: 30ish, despite identical results.

Honestly the more past results it takes into account the more it might solidify what it outputs. Though I could just be talking nonsense as you're the one who's running it not me.

Feel free to continue to work on it and post it, it's interesting (as someone who loves simulation games that are based on reality) and I'd be interested in seeing it take into account factors like form, fatigue, random resurgences, and other improvements.

I know it's not meant to be accurate - just more than a standard prediction it feels more like saying 'algorithmically your performance is poor' with those nations it predicts out while the biggest part of its considerations is past performance.

(to be transparent, my form has been pretty good this year so diverse support is my explanation for doing badly on it, though if it is random then maybe it's just me being unlucky)

2 hours ago, diva thin muffin said:

@Iz 🌟

I actually already do add a little randomness factor. As for your assumption, I'm afraid I don't understand what exactly you are saying to be honest. For what it's worth, I'm not trying to claim that my prediction is accurate in any way. It didn't even manage to accurately predict my own votes. My top score wasn't even in my predicted top 10 lmao. As for your last comment, I'm not sure what to say but I've already tweaked the algorithm quite a bit between months, and the data I'm using goes a lot further back than just the past year now.

For clarification, it takes into consideration:

• Points given in total vs Max possible points in the same time period

  • In other words, if Abcada has voted in 3 finals and 2 semi finals where Effeghe was competing, but only gave 3 points in total, their score would be 3/(3*18 + 2*12) = 3/(54+24) = 3/78 or 3.85%

• Additionally, I also take into consideration how often country A gives country B any amount of points vs how often it has had the opportunity to do so.

• Finally, I add a little randomness factor to get a little variety.

Hope that answers at least some of your questions :) but again, this is not meant to be accurate. All it does it allow me to make a prediction without my personal bias showing lol. And yes, there are obviously some patterns that will show up, since some people do very consistently vote for the same countries but then again some countries also do very consistently do well in the contest so maybe this pushes people to try to be aware of their own biases when voting.

But if this upsets people I can certainly refrain from posting it again and just keep it to myself.

Brb changing my country name to Effeghe