Skip to content

Building a Better Fish Party

  • Alex 

This article was written in November 2016 when I was Head of Product for the poker and bingo business unit at Microgaming.

When we launched Fish Party SNGs earlier this year, we introduced the industry’s first progressive jackpot to Jackpot SNGs. The problem is, when you’re first to do something, to a certain extent you’re flying blind. You probably won’t get everything right first try.

Fish Party SNGs are a great example of that. When we first set out to design the game, we had ‘go big or go home’ in mind. So we created a game with a big initial jackpot, and big jackpot contributions, resulting in big prizes (as I write, the €1 jackpot is over 52,000 times the buy-in). To counter the large jackpot contribution, we made the operator margin very competitive – just 5%. But it wasn’t long after launch that we started to hear complaints.

Problem is, the size of our jackpot contribution – 5.17% – was so large that when added to the operator margin of 5%, the effective rake on a single Fish Party SNG became 10.17%. Some players complained that this made it impossible to win, especially considering that Fish Party SNGs are super-fast, designed to last just over 5 minutes. My immediate reaction to the complaints was simply ‘we designed the games with recreational players in mind’, but after a little more reflection I thought ‘what if these guys are right and it’s not possible to win’?

I believe that for a poker game to be successful, there needs to be a proper balance of luck and skill. Poker is not like roulette or slots, and players who play poker do so with different motivations. In any gambling game you need to deliver winning experiences frequently enough to keep players interested, but in poker the game must be beatable by some in the long term, or it’s not poker and a significant number of poker fans won’t play.

Given that our initial maths model for Fish Party was just an educated guess, our new Market Insights team and I set out to find out how we could improve it.

The first thing we realised is that it was indeed difficult to win – so difficult in fact, that very few skilled players were winning unless they somehow got lucky and hit a big multiplier or the jackpot. At this point we recognised that the feedback we’d received was correct. When even great players can’t win in the long term, the lifetime value of the whole player base is reduced because fewer games start. So we knew we needed to change.

But what change to make? To find out, we first identified those Fish Party players with the highest lifetime value, and plotted this against ROI to see if we could find a ‘sweet spot’, where players within a particular ROI range had a highest value.

We then created lots of new maths models, making one small configuration tweak at a time and simulating the effect on player ROI. The goal was to create a model which moved as many players as possible into the ROI sweet spot that we had previously identified.

Compared to the original maths model, the new model is more bottom heavy. Players will hit the lowest (2x) multiplier less often, but hit a 4x or 6x multiplier more often (the 4x multiplier actually replaces the 3x multiplier of the current maths model).

This results in a much improved ROI for skilled players, along with a lower probability that you’ll have the dreadful experience of hitting the bottom multiplier repeatedly. The only real downside is that the average jackpot size will decrease, because the progressive jackpot is increasing less quickly.

The new maths model will be rolled out on 14th December, subject to regulatory approval. Nothing will change except the maths model, and all of the current jackpots (some of which are massive!) will be carried over to the new game.

We will continue to monitor and tweak the game in the future, based on what our data tells us, and there are more Fish Party thrills to come in early 2017, as we add the game to our mobile and desktop browser platforms.

More Fee Tweaks

Since earlier this year we’ve been under a lot of pressure from our poker rooms to add fees to rebuys and add-ons in tournaments. There are many arguments for and against this change, some of which are flaky and some of which are undeniable. Two which are definitely true are:

  • Longer tournaments cost more to host. Bandwidth and hosting are two of the largest costs to any poker site. The argument is therefore that we should charge more for longer sessions – and rebuy tournaments are typically longer than the freezeout equivalent.
  • The more money is put into play, the higher the cost (in processing fees, risk, compliance etc). The argument is that because rebuy tournaments put more money into play they should be charged more.

During September’s Universal Championship of Poker, we added fees to rebuys and add-ons by accident (this is easier than it sounds, our tournament creation tool ‘helps out’ by automatically adding a 10% fee to all buy-ins, rebuys and add-ons, and this is easy to miss). We took the opportunity to make an experiment of the mistake, which unfortunately proved inconclusive.

We intend to run a proper experiment in January 2017, whereby fees will be added to a selection of the MPN’s rebuy tournaments. Each tournament will state clearly in its description that it is part of the experiment, along with the price of a rebuy and add-on. You can help us out by sending any feedback you have about this change to your poker room. We will take everything into account before deciding whether to implement such fees permanently.

The counter to the above arguments is of course that shorter games should be charged less. I agree with that, which is why at the same time we will reduce the fees in Super Turbo scheduled tournaments.


  • On December 14th, we’re reducing the effective rake on Fish Party SNGs from 10.17% to 5.85%.
  • In January, we’re going to experiment with fees on rebuys and add-ons in a selection of MTTs.
  • At the same time, fees on Super Turbo MTTs will be reduced.